yeah, I think my code is only interesting as a start to show how I ported it to circuitpython
#circuitpython-dev
jaunty juniper
jaunty juniper
ah you are using pyportal matrix protal, now I have a version that works on the ESP32S2, but I have to modify it to use with a wifi coprocessor. I want to do it because I'm interested in it, but I gotta solder headers on my airlift featherwing first 
jaunty juniper
if you are interested, I think I have a version that should work for you. I put it there:
https://github.com/Neradoc/websockets-circuitpython-test/tree/airlift
astral seal
if you are interested, I think I have a version that should work for you. I put ...
its works thank you
jaunty juniper
its works thank you
nice ! it still needs work though, it might be a little rocky
minor glade
hey, would y'all accept PRs on circuitpython libraries adding PEP-484 type hints?
manic glacierBOT
idle wharf
hey, would y'all accept PRs on circuitpython libraries adding PEP-484 type hints...
Ask again on Monday US time when the Adafruit crew are working.
fossil gorge
hey, would y'all accept PRs on circuitpython libraries adding PEP-484 type hints...
You could always ask by opening a Github issue. If no, they it's answered and closed, in a public, and helpful, forum. If yes, then you already have an issue from which you can create the PRs
lone sandalBOT
hearty tapir
@alpine nimbus so the crash happened in the requests code and not the Wi-Fi?
onyx hinge
hey, would y'all accept PRs on circuitpython libraries adding PEP-484 type hints...
I don't see why not, but there could be compatibility or tooling problems I don't know about. Why not start with just one, and we can look at whether it's working with mpy-cross, the documentation and so forth. Some type hints seem to work (def f(x: Optional[int]=None) -> int: ...) and generate a .mpy file of the same size as without the type hint. However, you can't import typing so Arbitrary generic types as base classes don't work, or any other technique that requires import typing. That's what I know.
Also, hi and welcome!
minor glade
I don't see why not, but there could be compatibility or tooling problems I don'...
ooh, that's interesting. i generated a stubs file for a few the libraries that had no types, and i was going through and adding types for the bits i was using. but then i thought, i'd need to do this after every update of the libs and it would be cool if the original stubs generation would be able to take the types from the file.
okay, is there somewhere i can learn how the libraries (like adafruit_midi etc) are built into the bundle and mpy files, so that i can try one and see if everything continues to work?
minor glade
Also, hi and welcome!
hi and thankyou!
onyx hinge
generally speaking, each lib in the adafruit bundle is its own git repo, so you'd want to fork that one and make a pull request there. Once the pull request is accepted, the next steps are that adafruit peeps make a release of the lib, then the bundle picks up the new tagged release, and it'll go out in a dated bundle zip
so for instance the magtag lib is at https://github.com/adafruit/Adafruit_CircuitPython_MagTag/ and improvements can be PR'd there
onyx hinge
In your Pull Request, github actions will automatically do things like build the mpy files, so you can just create the PR and check the status. Let me see if I can find some relevant guides..
solar whale
@minor glade If you have not seen this guide, it may help https://learn.adafruit.com/contribute-to-circuitpython-with-git-and-github/overview
onyx hinge
OK this is a bit out of date... https://learn.adafruit.com/creating-and-sharing-a-circuitpython-library/creating-a-library#mpy-2982472-11
minor glade
ah neat, thank you :)
i have a few of the libraries cloned already, so i'll read their github actions to see the build and test workflow and then make some changes and see what happens
onyx hinge
OK, I did a quick update on the Learn page I linked, some of the version numbers it stated were out of date.
@minor glade feel free to ask more questions if you get stuck. I'm stepping away for the moment though
blissful pollen
@minor glade one gotcha I ran into first contributing to the libraries is the CI runs black and pylint on all the code and won't accept it unless it passes those checks. If you can't figure it out let me know I can probably help you through
minor glade
<@!589228615812775972> one gotcha I ran into first contributing to the libraries...
ah, i just recently got my editor set up to run Black on save, i'll see if i can teach it about pylint ^_^
manic glacierBOT
onyx hinge
this asm trick for function-specific ABI is horrifying. I'm sure someone motivated could use it to make CircuitPython compile smaller and be less readable. https://justine.lol/cosmopolitan/index.html
(different for arm than x86_64 of course)
manic glacierBOT
naive cave
Is there a Board that can run circuitpython and be programmed via Bluetooth? As far as i can see, Circuit Playground Bluetooth etc. Can't, at least out of the Box...
I guess that would need either a custom Software in the Host, or a file-access bt-profile.
Alternatively, is there something like kind of "USB over bt" that could be added externally, transparent to the board?
manic glacierBOT
hearty tapir
@atomic summit is this a DEBUG=1 build?
hearty tapir
@analog bridge I think that the same pins are not supposed to be used twice, thus the hang.
analog bridge
<@747673663419187281> I think that the same pins are not supposed to be used twi...
hi @hearty tapir, I don't think that is the case... I did a fresh start, I am using different pins in order to not get the in use error.
analog bridge
I think I found the issue... common_hal_reset_pin(NULL) causes the crash.
onyx hinge
excellent, thank you.
tulip sleet
@slender iron I had to push to your PR via the web. I am curious, if you mouse over question mark on this permission checkbox (right sidebar on the PR), does it mention adafruit/circuitpython or micropython/micropython? (This screenshot is from an old draft PR of mine.)
I wonder if it's different because you forked from micropython
turbid radish
?serverinfo
digital shoreBOT
adafruit
onyx hinge
Hi all and particularly <@&356864093652516868>, here's your reminder that due to the holidays there's no meeting today. I do want to take a quick moment to thank the past week's authors and reviewers, though!
- 27 pull requests merged
- 17 authors - siddacious, digixx, microDev1, bettse, ladyada, jamesrusso, slootsky, makermelissa, mshannon78660, d-c-d, ktdreyer, TwoTenPvP, dhalbert, jay0lee, jepler, twa127, jerryneedell
- 9 reviewers - siddacious, jerryneedell, ladyada, tannewt, brentru, makermelissa, dhalbert, FoamyGuy, jepler
- 16 closed issues by 11 people, 19 opened by 18 people
There are lot of people here, and a bunch of names I'm not familiar with: digixx, bettse, slootsky, jamesrusso, d-c-d, ktdreyer, TwoTenPvp, jay0lee, and twa127. We appreciate all your contributions!
We have one new library, https://github.com/adafruit/Adafruit_CircuitPython_LTR390
Thank you all for a great 2020 and we'll talk to you in 2021! This community has been truly important to me in this difficult year, so treat yourselves to a bug hug report from me.
lone axle
I'm going to post some hug reports for the week here:
- jepler: Thank you for working on a fix for Acenders / Decenders and fonts without the M Glyph in display_text
- makermelissa: Thank you for the Web ESPTool. It's very convenient to use this instead of the python tool with drivers and extra setup steps.
- Group Hug: wishing everyone a happy new year!
cobalt grail
I want to give myself a big hug for the swedish translation (Not sure anyone else uses it to be honest). 😎
Also; a big distanced group hug to everyone!
lone axle
🎉 💯
crimson ferry
Do I need the ARM toolchain for building esp32-s2?
tulip sleet
no, when you run esp-idf/install.sh, it will download its own toolchain copy.
crimson ferry
I'm looking at the Building CircuitPython guide, but sounds like I should start with Espressif's getting started guide
tulip sleet
@crimson ferry take a look at ports/esp32s2/README.md, which will then point you elsewhere as well. Basically, you do esp-idf/install.sh once, and then after that, whenever you want to set up for building, do esp-idf/export.sh in a fresh terminal.
blissful pollen
@crimson ferry I set up up building esp32-s2 not too long ago so if you get stuck or run into some weird problem let me know I may remember. It didn't end up being that hard to do
crimson ferry
OK, thanks, I've never built any software before, so I'm trying to get my arms around what all's needed. I'll start there.
tulip sleet
they've made it a lot easier than it was for the 3.x esp-idf versions. Setting up to build nina-fw was a lot more painful
slender iron
Is there a Board that can run circuitpython and be programmed via Bluetooth? As ...
Not yet, but it's something I'd really like to enable for all of our nrf products
slender iron
<@!252717193496756235> I had to push to your PR via the web. I am curious, if yo...
It says adafruit/circuitpython for me
crimson ferry
is it normal to get esp32-s2 install.sh errors like ... ERROR: No matching distribution found for bidict>=0.21.0 (from python-socketio>=1.6.1->Flask-SocketIO<3.0,>=2.9->gdbgui==0.13.2.0->-r ...
maybe related to Python version, some warnings eearlier about Python 2.7
tulip sleet
@onyx hinge the ESP32-S2 I2S PR https://github.com/adafruit/circuitpython/pull/3733 has some unresolved comments. I can pass it to someone to test, but I though there were some loose ends to wrap up.
@crimson ferry python should be python3
crimson ferry
wow, even on Big Sur it's not
tulip sleet
i have NOT tried to do ESP32-S2 builds on Mac. Only Linux
slender iron
I think I did S2 builds on mac
crimson ferry
there are apparently a lot of wrong ways to make macOS use python3
tulip sleet
ideally ESP-IDF should refer to it as python3, not python, but it doesn't...
jaunty juniper
I use docker and ubuntu to build on mac
crimson ferry
would I be better off having a dedicated Raspberry Pi for this?
tulip sleet
it would be really slow
slender iron
I wouldn't. the pi builds slowly
last flame
alias python => python3... I've used latest IDF on macOS. No need to build anything just clone, checkout s2 branch, recursively pull in submodules, run the ./install.sh and . ./export.sh commands and ya should be set.
ornate breach
Is there a good learning guide on using busio register for CP?
I am trying to verify an issue of always getting back 255 from register reads on arduino and I want to test on CP, but I haven't built the CP library yet
this is for a breakout board I am working on
crimson ferry
@last flame the git stuff is still a foreign language to me, but don't I want to work on main?
slender iron
@crimson ferry ya, you want to start with main
tulip sleet
but if you are making changes, create a branch off of main to keep your changes separate. Don't commit to main.
crimson ferry
I did make a branch of my clone of my fork
tulip sleet
we renamed master to main
last flame
oh magic
tulip sleet
or more correctly, we made a new branch main and abandoned working on master
main is now the default branch
blissful pollen
Is there a good learning guide on using busio register for CP?
I've spent a lot of time poking at this area for the work I'm doing so I can try to help
ornate breach
oh cool, you can dm me so we don't clog this up 🙂
slender iron
please don't
this is a better place to discuss technical things
because then others can lurk and learn
ornate breach
@blissful pollen C code: Serial.print("The device ID is: "); Serial.println(ODT_AT42QT1070().readRegister8(0x00)); // shoule be 0x2E delay(1000); returns 255 rather than 46
ornate breach
well i need to verify in CP
so I need to write a quick CP program to fetch the register value
but I think i found it
ornate breach
Okay, so it works on the circuitpython side by I'm not understanding why the arduino side isn't... hmmm
manic glacierBOT
tulip sleet
@slender iron the 6.0.1 circuitpython-org action failed. I'm looking at it but you changed it in main (maybe to fix a problem). Does this error ring a bell?
Run python3 build_board_info.py
python3 build_board_info.py
shell: /bin/bash -e {0}
env:
pythonLocation: /opt/hostedtoolcache/Python/3.8.6/x64
RELEASE_TAG: 6.0.1
ADABOT_GITHUB_ACCESS_TOKEN: ***
Logged in as adafruit-blinka
Traceback (most recent call last):
File "build_board_info.py", line 310, in <module>
generate_download_info()
File "build_board_info.py", line 257, in generate_download_info
previous_languages.update(version["files"].keys())
KeyError: 'files'
Error: Process completed with exit code 1.
https://github.com/adafruit/circuitpython/runs/1618736022?check_suite_focus=true
slender iron
@tulip sleet ah, ya. I removed the files part from the json
since it was too big to download
tulip sleet
so it's having trouble with the existing json that's there. I can maybe patch up the existing one and add the download counts
not sure it's worth writing some code for upward compatibility
slender iron
you can run the script locally and do it manually
tulip sleet
yes, I can run the main version on the 6.0.x branch
tulip sleet
got it, tnx
tulip sleet
np, it's an edge case
tulip sleet
@slender iron could you review ^^? It's set so I can't self-review. Thanks!
slender iron
looks
slender iron
@tulip sleet I'll let you merge
manic glacierBOT
tulip sleet
manic glacierBOT
Using the following test code and a SGP40 attached to the CLUEs Stemma QT port,
import board
import busio
import adafruit_bus_device.i2c_device as i2c_device
i2c = busio.I2C(board.SCL, board.SDA)
i2c_device = i2c_device.I2CDevice(i2c, 0x59)
print("i2c_device 'found'")
with 6.0.1 I get:
with 6.1.0 Beta 2 I get:
 • I see a lot of "make ... flash" but that makes no sense to me, the binary should be independent of USB port??
Also, how did you get around the python2 vs python3 thing on macOS? I went down a deep rabbit hole today on pyenv vs venv vs virtualenv vs etc etc, and came out not wanting to do any of that.
idle wharf
Are you going to build all the ports or just esp32s2
idle wharf
You don't need to worry too much about virtualenv because when you build you'll be in a virtulenv supplied by the IDF.
However my $.02. Use pyenv on Mac. SOO much easier IMHO than other methods. 😉
I did nothing in relation to the case-sensitivity of my volume.
Make is just a command that runs scripts which build the binaries, along with the MAKE command you can pass additional parameters like BOARD=name_of_board and you can also pass clean which cleans up crud from a previous build or flash wich will flash your device so you don't have to think about what the 20 parameter esptool command is
But you can just run MAKE BOARD= and do the flash step manually
crimson ferry
I thought that might be the case, flashing to me is an entirely separate thing. Still not sure if I can get a UF2 out of this process, or a .bin
idle wharf
when I look at a recent build output, I only see a .bin file
crimson ferry
cool, I won't worry about UF2 yet
idle wharf
I've not played around with the UF2 because the build flashes my device for me... so why bother. 😉
But it is the way most other people will install so its good to get to that method at somepoint.
crimson ferry
makes sense. So pyenv was how you got install.sh to play with python3?
analog bridge
make BOARD=... outputs a .bin which is automatically converted to .uf2
idle wharf
run python --version in your terminal
idle wharf
`make BOARD=...` outputs a `.bin` which is automatically converted to `.uf2`
Knowledge! I didn't know that!
crimson ferry
2.7.16
idle wharf
yes.
crimson ferry
".bin which is automatically converted to .uf2" not sure what that means... they operate differently, no?
idle wharf
I had to think about. I use Pyenv to control which Python environment is active. The System one is the default.
* system (set by /Users/askpatrickw/.pyenv/version)
see the makefile... there uf2conv.py (converts to uf2) is run after esptool.py (generates .bin)
crimson ferry
ok, thanks, I'll check it out
analog bridge
have you done this step to make python3 default?
https://docs.espressif.com/projects/esp-idf/en/latest/esp32s2/get-started/macos-setup.html#installing-and-setting-up-python-3-as-default
crimson ferry
I infer those sections at the end correspond to make params
Big Sur doesn't allow even sudo to mess with /usr/bin
but looks like that is local, so I'll dig into that more
idle wharf
The setup stuff is annoying and then you don't need to do it again. You'll get it.
crimson ferry
Thanks for all the help @idle wharf @analog bridge I'm still at the stage where every little thing requires a bunch of searching and reading, so sanity checks and even small tips are immensely useful
crimson ferry
have you done this step to make python3 default?
https://docs.espressif.com/proj...
I can create /usr/local/bin/python -> /usr/bin/python3 but I can't create a symlink for /usr/bin/python, even sudo, and install.sh still fails. /usr/local/bin:/usr/bin: is the order in $PATH. Any ideas?
analog bridge
🤷♂️
crimson ferry
ok, thanks, i've got a few other avenues to try
idle wharf
I'm curious let us know what you try and what works I've been putting off upgrading because I heard someone had trouble with the IDF on Big Sur. I can remember who it was..
🔍 Searching for "Big Sur" reveals BruceS and Jerryn have it working...
analog bridge
question regarding esp32s2... which adc channel has interference problem when WiFi is on?
last flame
interference? no idea... but:
Since the ADC2 is shared with the WIFI module, which has higher priority, reading operation of adc2_get_raw() will fail between esp_wifi_start() and esp_wifi_stop(). Use the return code to see whether the reading is successful.
For a given channel, adc2_config_channel_atten() must be called before the first time this function (i.e. adc2_get_raw) is called. If Wi-Fi is started via esp_wifi_start(), this function will always fail with ESP_ERR_TIMEOUT.
https://docs.espressif.com/projects/esp-idf/en/latest/esp32s2/api-reference/peripherals/adc.html
not sure how that fits into the CPY code 😬 would have to go source-diving
analog bridge
Thanks! @last flame
solar whale
@crimson ferry FYI -- in my experience, you only need a case sensitive file system on your Mac if you need to build the xtensa toolchain. I am able to build CP for the esp32s2 on a a standard file system following the guide for installing the esp-idf.
jaunty juniper
so, I have run into a problem on macOS with circup that is a problem with shutil.rmtree that I actually encountered earlier when doing my own idea of circup. The problem is that rmtree gives a FileNotFoundError when deleting a file that has a matching ._ hidden file. Or more exactly it seems to properly delete the file, but then errors on the hidden file. I originally thought of posting an issue on circup, with possible solution of adding ignore_errors = True (which might still be a valid workaround), but maybe it's a problem with Circuitpython/tinyUSB ? (or yeah with macOS, but...)
>>> import shutil
>>> shutil.rmtree("/Volumes/CIRCUITPY/lib/adafruit_display_text/")
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/Applications/Xcode.app/Contents/Developer/Library/Frameworks/Python3.framework/Versions/3.8/lib/python3.8/shutil.py", line 715, in rmtree
_rmtree_safe_fd(fd, path, onerror)
File "/Applications/Xcode.app/Contents/Developer/Library/Frameworks/Python3.framework/Versions/3.8/lib/python3.8/shutil.py", line 672, in _rmtree_safe_fd
onerror(os.unlink, fullname, sys.exc_info())
File "/Applications/Xcode.app/Contents/Developer/Library/Frameworks/Python3.framework/Versions/3.8/lib/python3.8/shutil.py", line 670, in _rmtree_safe_fd
os.unlink(entry.name, dir_fd=topfd)
FileNotFoundError: [Errno 2] No such file or directory: '.___init__.mpy'
>>>
(the __init__.py file has been deleted)
crimson ferry
thanks, @solar whale
analog bridge
just a reminder... issue #3886 needs to be closed manually... the link to link issue and pr isn't appearing for me.
onyx hinge
I don't think esp-idf requires that python be python3, I have it as 2.7 on all my machines and circuitpython builds for esp32-s2. I would investigate other theories before continuing with the python-is-python3 theory @crimson ferry
@jaunty juniper Looks like various people have encountered similar problems in other contexts. https://bugs.python.org/issue29699 https://bugzilla.redhat.com/show_bug.cgi?id=1362667 etc -- I think that macos is manipulating the ._ file for you, and it results in this problem readily. A fix has been offered to Python but not incorporated yet .. https://github.com/python/cpython/pull/14064
https://docs.python.org/3.8/library/shutil.html#shutil.rmtree you might be able to specify ignore_errors=True ?
crimson ferry
@onyx hinge thanks. the error I'm getting is basically due to requirement bidict>=0.21.0, and bidict docs say "Drop support for Python 2" as of bidict 0.19.0. Espressif instructions also step through making python3 the default python (though I couldn't get this to work on Big Sur) ...working on pyenv now
IDF version 4.2 does not have the python3 instructions, but it's in the latest, but the error occurs invoking install.sh from the current circuitpython main
onyx hinge
oh yeah it's possible some upstream lib has ruined things since the last time I installed esp-idf
crimson ferry
is there a way to copy circuitpython Actions to a forked repo? i realize local would be faster dev cycles, but I have diminishing confidence that anything I build locally would match what GitHub builds
analog bridge
is there a way to copy circuitpython Actions to a forked repo? i realize local w...
yes... they can be enabled from the actions tab
onyx hinge
I have done that for my repo, it works.
crimson ferry
I thought I did that, but my Actions are empty
I'll start over and re-fork
this results in no workflows in the Actions tab
or is that supposed t be empty until I commit or merge something
ok, nevermind, it needed a push
can I make DEBUG=1 this way though?
crimson ferry
I have it installed, trying to decide whether to install new python into its directory, or symlink to existing
idle wharf
you use pyenv to install and set which python version is active
I would not try to make it work some other way... seems like more potential for issues
crimson ferry
ok, no symlink then, thanks
idle wharf
You could setup the GitHub actions, keep in mind, it will try to build the world and there are Adafruit specific config and secrets so you'll need to modify the github actions to be your own.
crimson ferry
ok, so that's likely a dead end (along with the DEBUG issue to see logging on the debug console)
idle wharf
That would be one of the changes you'd need to make 😉
crimson ferry
do we have a preferred version of python3 for the build?
idle wharf
So funny thing... you're just fixing this so the install.sh will work. When you start the IDF environment.... it has its own Python version in its own virtualenv
crimson ferry
ok, so I don't need to worry about that, proibably any python3 will work
idle wharf
I use 3.8 for system
slender iron
@crimson ferry I think I used brew for python3
idle wharf
Did you use ln to set /usr/bin/python to python3 ?
He can't make that change on Big Sur
slender iron
looks
crimson ferry
pyenv python3 build failed, and myriad different "here's how I fixed it..." search results o_O
I could just buy a Linux box
crimson ferry
yeah, but Big Sur changed some things
slender iron
ya, I'm trying it now
I've building another port but not esp
I'm trying to get esp going now
(and I'm on big sur)
idle wharf
what do you get from python --version on a new terminal ?
crimson ferry
2.7.16
slender iron
same
idle wharf
Kewl... I was worried maybe you had tweaked your OS and then upgraded. You're both starting from the same state it seems
slender iron
I have been able to build other ports. I just haven't tried esp on my mac recently
I do have the repo on a case sensitive partition
slender iron
what do folks think about doing a 6.1.0-rc.0?
I think it'd be good to do a stable release before we change the socket stuff to support tcp server
lone axle
@onyx hinge do you have any thoughts or plans on a library in the community bundle perhaps that includes your icons.py mapping for fork awesome? Maybe a few other helpers as well like a function to return a single character label or button ready to add to the screen could be nice as well.
analog bridge
@slender iron I am having some trouble because of idf@v4.2.... the latest one is with touch-alarm... I am getting an error deep inside the idf
https://github.com/microDev1/circuitpython/runs/1620571067?check_suite_focus=true#step:12:850
slender iron
<@!252717193496756235> I am having some trouble because of `idf@v4.2`.... the la...
That seems like the BIT macro may be wrong
I'd rather not rush things. we can do 6.2.0 betas
making 6.1.0 stable will mean we can recommend the same version for everyone
lone axle
I am in favor of 6.1.0-rc.0 would be great to MagTag onto the same stable release as everything.
slender iron
looks at the 6.1.0 issues
analog bridge
idf@v4.2 is blocking lots of stuff reset-to-bootloader, touch-alarm and ulp-alarm plus some upgrades for dualbank module...
slender iron
@analog bridge I don't want us on some random idf version
we can patch our branch if needed or simply not use the idf for things
slender iron
@idle wharf do you think #3735 should block 6.1.0?
onyx hinge
@lone axle It's worth considering!
slender iron
@onyx hinge how do you feel about us getting 6.1.0 stabilized?
onyx hinge
@slender iron I'm in favor of it 🙂 I just may not have been doing my fair share of the actual work.
slender iron
you could do the release 😛
onyx hinge
sure, I can do the steps
analog bridge
`idf@v4.2` is blocking lots of stuff `reset-to-bootloader`, `touch-alarm` and `u...
I guess these will have to wait then... I'll update the status on github.
idle wharf
<@!624005998696333331> do you think #3735 should block 6.1.0?
No, its an ESP32S2 only issue
slender iron
we could have an idf 4.3 tracking issue @analog bridge
onyx hinge
@slender iron OK, should I pick that up next time I'm ready to switch tasks then?
slender iron
@idle wharf I'd like 6.1.0 be stable for esp
@onyx hinge when do you think that will be? I can find time to do it today or tomorrow
idle wharf
<@!624005998696333331> I'd like 6.1.0 be stable for esp
OK... let me test is again
slender iron
it looks a bit weird calling connect multiple times
idle wharf
When should one setup Wifi.. inside or outside your While Loop? (Assuming no use of deepsleep)
idle wharf
Then let's close it or convert to a learn guide\doc bug.
slender iron
what happens if you do it that way? does it crash?
onyx hinge
@slender iron looks like I will probably spend the rest of my hours today wrapping up other stuff, so I could pick it up first thing tomorrow.
slender iron
@onyx hinge ok, that'd be great!
onyx hinge
will do
slender iron
thank you!
onyx hinge
a nice way to wrap up the year!
(I will be around but not "working" thursday through sunday)
slender iron
ya, that's fine. that's why we can do it as an rc
crimson ferry
@idle wharf esp32s2 probably could use a check in connect for already connected, otherwise it does seem odd to call it again
onyx hinge
@lone axle do you think that python module with the icon "names" is useful, then? I wasn't sure
@slender iron wait, will I make 6.1.0 or 6.1.0-rc.0?
slender iron
rc.0
idle wharf
what happens if you do it that way? does it crash?
Connect inside the loop, it doesn't crash the ping just doesn't work on the second time. It returns None.
Connect outside the loop it works. I agree with @crimson ferry though, connect should be more resilient.
lone axle
@onyx hinge I do think so for sure. I am used to using them in HTML/CSS by applying classes with their names. I'd love an analogous way in circuitpython that allows us to use them by name easily.
slender iron
@crimson ferry circuitpython built fine for me
crimson ferry
@slender iron are you using virtualenv, pyenv or the like?
manic glacierBOT
crimson ferry
install.sh is what fails for me, on a library dependency that seems tied to using python2 instead of python3
I'll start from scratch, slightly better informed this time, and see what happens
crimson ferry
what is the failure?
slender iron
hrm
I may have installed it myself
let me see what my python3 is
ya, I have python3 from brew
do you have it installed?
slender iron
k, I have 3.7.8 in /usr/local/bin
I also have 3.8.2 in /usr/bin but am not using it
crimson ferry
I have /usr/local/bin/python -> /usr/bin/python3 but it apparently doesn't get used despite $PATH order: /usr/local/bin:/usr/bin:
slender iron
my python itself is 2.7.16
ornate breach
Does the Qt Py build support bytes and bytearrays?
crimson ferry
@slender iron thanks for digging in, I'll start again as clean as I can and see what happens
ornate breach
I'm getting an error for unsupported datatypes
slender iron
looks like the install script from the idf did use 2.7
ornate breach
>>> import time
>>> import busio
>>> import board
>>> import odt_at42qt1070
>>> i2c = busio.I2C(board.SCL, board.SDA)
>>> qtouch = odt_at42qt1070.AT42QT1070(i2c)
>>> qtouch.touched()
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/lib/odt_at42qt1070.py", line 163, in touched
TypeError: unsupported types for : 'bytearray', 'int'```@ornate breach does it work on M4?import builtins dir(builtins)
ornate breach
guess not on qt py, let me test on m4
slender iron
@ornate breach what is on line 163?
that looks like the error you get when trying to add two different types
ornate breach
return self._buffer & 0xF
slender iron
and _buffer is a bytearray
you can't bitwise the bytearray
you need to [] an element of it
ornate breach
bummer, thanks
analog bridge
the error with touch-alarm isn't coming up with current main even though the include paths are the same
slender iron
@analog bridge you may need to include something before you include it
if its includes are wrong
ornate breach
oh I was attempting to mask the buffer but returning the individual bit is probably far more precise in the long run.
I wasn't sure if I wanted to leave it open to return the whole detect register or not.
strange pike
Does the Qt Py build support bytes and bytearrays?
qtpy: bytearrays yes, not sure about bytes
ornate breach
it's all good, it works now 🙂
slender iron
it should have both
ornate breach
just had to return the individual bit element
strange pike
are the official cirpy 6.0.1 downloads being built w/ the newer gcc-10-2020-q4?
crimson ferry
@slender iron bidict is apparently a dependency for something in requirements.txt (I can post the relevant parts of the output if of interest) ...always fails in the same place
slender iron
are the official cirpy 6.0.1 downloads being built w/ the newer gcc-10-2020-q4?
@tulip sleet would know
slender iron
<@!252717193496756235> `bidict` is apparently a dependency for something in `req...
what version of the idf are you on?
tulip sleet
no, the 6.0.1 github action is still running with the gcc 9 compiler
crimson ferry
the version in main
tulip sleet
did you do esp-idf/install.sh?
crimson ferry
yes, ebe7784. @tulip sleet yes, that's what's failing. I'll post the relevant bits of the output
strange pike
no, the 6.0.1 github action is still running with the gcc 9 compiler
thx ... would you expect any issues w/ gcc-10? i had the newer installed for building 6.1.x and threw it at 6.0.1, no errors so far, but not sure what to expect.
crimson ferry
slender iron
thx ... would you expect any issues w/ gcc-10? i had the newer installed for bu...
nope. I'm using gcc 10 locally without issue
strange pike
nope. I'm using gcc 10 locally without issue
👍
slender iron
esp-idf install.sh output: <https://gist.github.com/anecdata/65172f5cfff2bbdaf98...
I'm on python-socketio==4.5.1
you could activate the idf virtual environment and install that version directly
fossil gorge
I just did some CircuitPython stuff on weblate, and I was wondering if there's a way that I can test the strings? Such as test scripts or routines, and how to get the various languages?
onyx hinge
@fossil gorge I recently merged in the pending changes from weblate, 3 hours ago. If that would have included your contributions, you can grab the zip for the board you want to test at https://github.com/adafruit/circuitpython/actions/runs/451062198 and load it as you usually load circuitpython. If you did it since that time, weblate hasn't created a new pull request yet (it delays up to 2 hours)
@fossil gorge .. thank you for working on translating CircuitPython!
fossil gorge
I did it in the last 1-1.5 hours, so on par with my general timing!
I'll try it out for now, and keep an eye on the next release
onyx hinge
okay. If you go on github and look for the PRs from weblate you can use that to test. of course, if you want you can learn to build circuitpython on your own computer but that is more trouble than it's worth for a lot of folks
fossil gorge
Thanks @onyx hinge . I'm good (for now) impatiently waiting for the next build, and figure ways to ensure the right string appears in the right location.
The ones I'm mostly concerned or uncertain about, are the ones that appear to be mid-sentence, or sentence fragments - no starting capital letter and no period.
But also, the consistency of the terms across the codebase and their use
My mom was/is a translator, so I have... opinions on translations! lol
onyx hinge
I agree 100% -- there are a number of "best practices" for making software translatable that CircuitPython doesn't do, or doesn't do consistently. Sometimes it's due to technical limitations, and sometimes due to general unawareness. I think there's a big opportunity for someone with the right skillset to look through the core and figure out how we can improve that
fossil gorge
Hardcode all the strings!!!
No problem with one branch per language with all the same code, right?
</smart-alec>
blissful pollen
To build the SAMD builds now I need the new GCC10 correct? I'm pretty sure I saw that pass by in the last couple weeks
onyx hinge
my background is some college-level french, plus at a previous job I converted a HUGE existing codebase to use gettext-style translation. the result was very unsatisfactory. The translation system we have is I think mostly coded by @slender iron but the existing messages and message structure date from before translation was a thing.
fossil gorge
To be serious, I don't really have any experience with best practices, setup or anything similar, but I'm always happy to help out with what I do know, and learn + share the rest
onyx hinge
so I have some idea of what best practices are supposed to be, but not as much real world experience as I'd like
fossil gorge
I've worked mostly in the MS product space, so my experience mostly revolves around resource files
onyx hinge
Thanks! Hearing from actual multilingual people is also helpful, when so many people at the center of development are only fluent in American
ornate breach
coughs
onyx hinge
hehe
ornate breach
I'm not sure I'm fluent in american and I AM america hehe
fossil gorge
So the incorrect AND correct spellings (in that order) of words 😉
ornate breach
I know enough spanish and french to get food and get myself arrested
onyx hinge
haha! I can order food in german and I can get myself in pretty good trouble in french still.
manic glacierBOT
ornate breach
I still have my textbook from my college french days
it was a downhill battle, B, C, and a D my final semester 🙂
onyx hinge
@fossil gorge ^ there is the pull request right on schedule. In an hour or so I think this page should have links to download: https://github.com/adafruit/circuitpython/pull/3893/checks
fossil gorge
Saw that. I think GithubBot is spying on the convo!
ornate breach
um, it was advanced french by the third semester
so mostly just actually putting it together. The instructor played a big part in this. She was of course a middle aged french lady who had fun making fun of english speakers
fossil gorge
Ugh, hate those types of teachers
Rather than elevate others, they make themselves feel better by putting them down
onyx hinge
I did okay right up until classes switched from grammar & pronounciation to "reading serious literature". I can't read serious literature in English.
fossil gorge
Students absolutely THRIVE in those environments
ornate breach
I can read okay in french but speaking, I can barely do that in english
fossil gorge
Same here, with the "Serious literature". I need to be interested, regardless of the language. And "a good grade" != "interest"
tulip sleet
To build the SAMD builds now I need the new GCC10 correct? I'm pretty sure I saw...
We are building with gcc10 (it might work with gcc9 but you should use 10 because it's stricter and will catch some things 9 wont'). Versions discussed here: https://learn.adafruit.com/building-circuitpython/linux#install-build-tools-on-ubuntu-2986713-2
ornate breach
I took a number of years of spanish growing up and did pretty well. not enough to contribute to society but yeah
fossil gorge
I did about 3.5 years of Spanish in high school, and now live in Texas. Pretty sure I can get away with enough to take a severe beating in the wrong crowds!
blissful pollen
We are building with gcc10 (it might work with gcc9 but you should use 10 becaus...
Thanks Dan, figured may as well get it all set up
fossil gorge
Specially combined with the Spanish I picked up from the movie "Encino Man"
crimson ferry
@idle wharf is there some reconciliation process that you did after installing the default Espressif IDF, but then circuitpython requiring what I think is an Adafruit-patched version of the IDF from some branch of Espressif's tree?
idle wharf
<@!624005998696333331> is there some reconciliation process that you did after i...
You pick which one you're using when you run the export.
alias get_idf
get_idf='. $HOME/esp/esp-idf/export.sh'
or if you run export from the ..ports/esp32s2/esp-idf folder
crimson ferry
@idle wharf ok, thanks, that's what I did
idle wharf
I ran export.sh from both and then did where python and they both seem to use the same virtualenv (in ~/.espressif/python_env/idf4.3_py3.8_env/bin/python) though.... so you could think your using one and because the virtualenv was created by the other end up with a mish-mash.
messy dependency mgmt for sure
crimson ferry
@idle wharf yeah, once this works, I may need to do another OS clean install to make sure it can be replicated... I should get back into VMs
The error I'm getting in install.sh is in Step 3 of Espressif's process. I didn't do step 2, but instead just used the IDF already within the circuitpython repo. Is this a bad assumption?
https://github.com/adafruit/circuitpython/blob/main/ports/esp32s2/README.md is high-level, but seems to support that sequence
idle wharf
I believe you should install.sh from the idf in the CP repo.
That will eliminate the virtualenv not being from the right instance of the idf.
(I prolly need to do that... )
I was planning to rebuild when the new GCC comes down
tulip sleet
i always use install.sh (once) and export.sh (in each new shell), and that works out well
gcc10 is not used for the esp32s2 builds - we use the toolchain that esp-idf installs
onyx hinge
hmm I didn't read the whole thread but this reminds me of our problems trying to reset wifi https://github.com/espressif/esp-idf/issues/5122
crimson ferry
6.1.0-beta.2-171-g8b03951fe on 2020-12-29 🇸 2️⃣ 🎉 I (106604) wifi:connected...
hearty tapir
`6.1.0-beta.2-171-g8b03951fe on 2020-12-29` 🇸 2️⃣ 🎉 `I (106604) wifi:connected...
congratulations @crimson ferry
crimson ferry
started over... brew installed python 3.9, but it works, I'll take it
stuck elbow
does https://learn.adafruit.com/building-circuitpython/linux need updating for gcc version for 6.x?
strange pike
does https://learn.adafruit.com/building-circuitpython/linux need updating for g...
there's a mention already there: "For CircuitPython 6.1 and later, we are now using 10-2020-q4-major."
stuck elbow
oh, I didn't see it, I'm blind, sorry
idle owl
Better to suggest it and have it there already than to have missed it in the first place.
strange pike
there are a couple examples (setting path etc, towards the bottom of that page) that still show the eabi-9-2019 version.
the version links probably ought to be listed explicitly on the Manual Setup page for clarity
tulip sleet
the version links probably ought to be listed explicitly on the Manual Setup page for clarity
I have clarified the version information on both the Linux page and the manual setup page.
strange pike
> the version links probably ought to be listed explicitly on the Manual Setup p...
brilliant, tysm
maiden cairn
Is there a separate channel for #adabox016? I couldn't find one. I am a Discord noob too so not sure if I am searching correctly. Anyway I have a question about the Matrix Portal. When the 64x32 RGB Matrix is connected to the Matrix Portal if I connect it to my laptop USB, I don't get a very stable connection and my laptop starts behaving weird. Is this a power draw issue? Should I disconnect the Portal from the RGB Matrix when programming it?
slender iron
@maiden cairn we don't have separate adabox channels. yes, that sounds like a power issue. The matrix should have a power source of its own when you are programming it over usb
maiden cairn
Thanks @slender iron . I just got around to trying out some of the other adabox016 examples and wanted to test the halloween countdown, but change the code to countdown to New Years Eve.
slender iron
good idea!
maiden cairn
I guess I will disconnect the RGB Matrix first. Not sure how I can power the matrix and also keep it connected via USB. There is only one USB C port on the Matrix Portal.
slender iron
@maiden cairn do you have a powered usb hub? You could try using that between your computer and the matrix
maiden cairn
@slender iron Aha! Yes! Thanks. Completely forgot that I could use something like that.
BTW, a bit unrelated question, but since the Matrix Portal has ESP32 and is connected by WiFi to my local network, is there an example code to connect to the REPL over-the-air?
slender iron
nope, we only support usb now
crimson ferry
esp32-s2: do we need start_station() in start_scanning_networks? scanning should be independent of wifi mode (e.g., scanning can be done unconnected or connected)?
manic dune
Does anyone recognize this error?: TypeError: unsupported types for ge: 'AnalogIn', 'int'
The program runs once, and the error stops it the second time through
while True:
psi = mpr.pressure * 0.0145037738
print("psi:", psi)
print(a0.value)
print()
time.sleep(1)
if a0 >= 65000:
relay.value = True
elif a0 < 65000:
relay.value = False
it stops at the if statement
hearty tapir
esp32-s2: do we need `start_station()` in `start_scanning_networks`? scanning sh...
Then it would default to AP_STA mode. https://github.com/espressif/esp-idf/blob/b0150615dff529662772a60dcb57d5b559f480e2/examples/wifi/scan/main/scan.c#L125 it is done like that also in the espressif examples. I think you asked this already, can you tell me the suspected issue you see? It may help me to better understand which issue you try to solve 🙂 Thank you
hearty tapir
esp32-s2: do we need `start_station()` in `start_scanning_networks`? scanning sh...
I haven’t worked on the PR to fix the Wi-Fi mode vs Wi-Fi interface yet, if you want to address this yourself - the issue you found :). I‘m on vacation and would usually be all over this type of stuff, but right now I spend time with family 🙂
ornate breach
question with the cp community bundle
the driver i'm adding has a release for it, but the CI check is saying it's failing because it doesn't appear to have a release yet
says it failed to generate a version file
fatal: run_command returned non-zero status for drivers/at42qt107
9.
10
11 Generating VERSIONS
12Failed to generate versions file. Its likely a library hasn't been released yet.
13WARNING: some failures above```how would I fix this?
tulip sleet
@ornate breach the commit for at42qt107 that you pushed is 49261b9, which is way back on the commit list for that repo. Your release is de255a8. So you need to descend into the submodule repo, check out main to the latest, and then go back up to the bundle repo and commit that change.
ornate breach
ah okay
tulip sleet
e.g.
cd at42qt107
git checkout main
cd ..
git commit -a -m "..."
you can verify you're at a tagged commit by doing git log
ornate breach
wait, I think i got it
I fetched and rebased in the submodule so hopefully that should work with a release I just created
yes it worked for that part
thanks @tulip sleet
🙂
tulip sleet
@ornate breach a rebase was probably no necessary, just checkout the main branch, because the most recent commit was the release. If that were not true, you can always check out a particular commit: git checkout abc12345
a submodule is a snapshot of a particular repo, at a particular commit
ornate breach
It passed the CI for the community bundle after fetching main so I think it’s okay for now
I appreciate the help in getting it sorted 🙂
Still new to these sub module shenanigans, but I’m slowly getting it down
tulip sleet
i have found this book helpful: I read it a number of years ago: https://git-scm.com/book/en/v2
crimson ferry
Then it would default to AP_STA mode. https://github.com/espressif/esp-idf/blob/...
I'm just trying to understand the design, and figure out what code is anticipatory of additional modes, and hopefully keep clean concepts of netif [de]init, wifi [de]init, wifi mode (we'll eventually want AP and AP+STA, I hope), wifi start/stop, wifi [dis]connect. Sounds like your PR does some of this.
crimson ferry
I haven’t worked on the PR to fix the Wi-Fi mode vs Wi-Fi interface yet, if you ...
Sure, I'll PR the start_station issue I found, good small test of the process. 🙂
prime cove
I have some code that is hard-faulting my board and not entering safe mode. Reflashing CP hasn't worked since it re-runs the offending code immediately. Is there a way to reset a board into the REPL from power up?
onyx hinge
@tulip sleet should I create the 6.1.x branch FIRST and then tag the -rc.0 SECOND ?
@prime cove yes but it depends on the device you have. Can we take this over to #help-with-circuitpython ?
ornate breach
So sometime after the New Years I (like a few people here) will be getting one of the ESP32-C3 boards
I’m hoping to contribute to support for that chip for CP since native USB is on it. It’s kind of exciting
I have no idea how much I’ll be able to do, but I’m excited for that
onyx hinge
hopefully very similar code will work on both since they're both able to use the esp-idf.
(as I understand it anyway)
ornate breach
Hopefully so
manic glacierBOT
onyx hinge
If you have items you think should be noted in the Known Issues section of 6.1.0-rc.0 please let me know. I'm drafting the release notes.
Known issues
- ESP32-S2: Crash when repeatedly creating and destroying busio.I2C object on ESP32-S2 (#3846)
- Built-in adafruit_bus_device has been disabled due to issues (#3859, #3856), please install the version from the Bundle
- See https://github.com/adafruit/circuitpython/issues for other issues.
solar whale
@onyx hinge are you planning to update the frozen libraries before releasing 6.1.0
onyx hinge
I'm not sure I know how to do that.
tulip sleet
i can do that -- I have a script
onyx hinge
It would make a PR?
@tulip sleet has also reminded me that there are two 6.0.x-tagged bugs https://github.com/adafruit/circuitpython/milestones/6.0.x that we should look at
tulip sleet
i make the PR by hand, but I have a script to do the submodule update
solar whale
Will bus_device be fixed before the release?
onyx hinge
no, use the python-coded bus device.
solar whale
hmmm -- should boards that used to "freeze" bus-device be updated now to put it back in?
tulip sleet
I think that might have happened when the PR was reverted, but I will double-check
onyx hinge
I did not revert the PR, I just disabled it in one of the .mk files
solar whale
ah -- ok -- thinking of m0_rfm69 and rfm9x
onyx hinge
so, no, if there are boards that froze it before and don't now, that would be a regression
blissful pollen
My goal is to start working on it today and try to get it working asap but it won't be today 😐
tulip sleet
OK, I will open a 6.1.0 issue for that.
onyx hinge
Thanks for raising these issues y'all
onyx hinge
Can someone point me at a board where this was changed? I am not finding it with git grep like I expect
(this = removing a frozen-in adafruit_bus_device)
blissful pollen
feather_m0_rfm9x should be one
onyx hinge
OK, found it -- thanks
commit 23ed3ef971f8f38497337874da39273e5ae090d7 Removing frozen libs
solar whale
@onyx hinge Thanks!
onyx hinge
you're welcome
slender iron
I’m hoping to contribute to support for that chip for CP since native USB is on ...
It isn't native USB. It's a built in USB to UART and JTAG
slender iron
we can support it once we have a BLE workflow
and it'll be BLE-only
because it can't do usb mass storage
ornate breach
makes sense, well if I can be of any help testing or contributing to BLE workflow I'd be happy to take a shot at helping move that forward 🙂
slender iron
adding ble support for it is the first step
onyx hinge
@slender iron should I go ahead and do the release we talked about but tag it as beta instead of rc? what would you like me to do?
onyx hinge
yeah there was a lot of stuff to cover when I started the release notes
we have a great community working on this thing!
tulip sleet
@onyx hinge thanks for doing the release notes, which can be tedious... I will work on updating the frozen libs right now
onyx hinge
@tulip sleet OK I'll wait for that, do a few release note updates, and tag there. Do we want to hold off with branching 6.1.x, or do it at the same time @slender iron ?
tulip sleet
@onyx hinge I forgot this! I had added a circuitpython/Makefile target make update-frozen-libraries, which does all the work.
onyx hinge
very nice
tulip sleet
so i'm using it
onyx hinge
I was thinking that some things should happen by a cron actions -- this could be one of them
"make translate" is the other one I think about
tulip sleet
it could be an adabot checking thing. Occasinally there have been reasons to delay an update
re 6.1.x: I wouldn't make a branch if we are still doing betas. An rc is feature freeze, then it's appropriate
onyx hinge
gotcha
tulip sleet
@slender iron @onyx hinge I would like to make a checklist of what's necessary for a release: there are many steps, and some pitfalls (like make sure you are tagging off the right branch). This could be in an internal Learn Guide, but I was wondering if we could put it in a wiki entry in the circuitpython repo. We have not enabled wikis yet. I started to do this and then wondered if it should be Learn Guide instead.
there are other examples of manual processes we should document. A README is not the best place, maybe
onyx hinge
I like the idea of a checklist, and I like the idea that it's in a place where people can contribute.
tulip sleet
though wiki entries don't have PR's, I think
anyway, something to think about, and we can always change it
idle owl
Issues can have checklists...
Markdown.
Wait I misread that
You want a general checklist, not a specific one.
Statement stands anyway.
tulip sleet
right, this would not be closed, though, so it's not an issue, but a checklist to refer to every time doing a release.
the wiki is markdown too, i'm sure
idle owl
We had an eternal library issue before. It's not unheard of. But fair enough.
Probably, yes.
tulip sleet
yah, an uncloseable issue is better as something else
onyx hinge
we could put the checklist in an issue template, and literally create a release issue with the checklist to check off.
idle owl
That's what I was picturing, I suppose.
tulip sleet
github has "projects" too, not sure how that works. Issue template, that sounds good. I forgot they existed (or that there could be multiple tempaltes)
onyx hinge
(however, issue templates are attractive nuisances, anyone could semi-accidentally create one. But then you just delete the unwanted text)
(especially if it's the only one)
ornate breach
submodules are a pain but i finally got the community bundle pr right
slender iron
my preference would be to leave wikis off. we can add something to the docs folder though
ornate breach
great now I mpy cross hehe... hey @slender iron can you look over this and see what might be causing mpy-cross to fail on this file. https://github.com/skerr92/at42qt-acorn-python/blob/main/odt_at42qt1070.py
ornate breach
doesn't have to be right now, but whenever you have a chance.
ornate breach
👍
strange pike
great now I mpy cross hehe... hey <@!252717193496756235> can you look over this ...
if you try importing in the repl you might get the same error unmangled. this is what i get:
>>> import odt_at42qt1070
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/lib/odt_at42qt1070.py", line 167, in set_lowpower
SyntaxError: 'break' outside loop
vs:
$ mpy-cross odt_at42qt1070.py
dexpresed ing back (most recent call File "odt_at42qt1070.py", line 167stionusEUSyntaxError: amcttsed ' outside l
thorny jay
Is there a new recommendation for MagTag and ESP32S2 board to use 6.0.1? https://learn.adafruit.com/adafruit-magtag-project-selector/code-the-magtag-project-selector say: "Make sure you have CircuitPython 6.0.1 or a more recent version (anything released after 12/28/20) installed. Previous versions including 6.1.0 betas will not work." ( cc: @trim elm )
tulip sleet
6.0.1 doesn't support magtag, so that needs to be rephrased
thorny jay
Yeah, that sentence was disturbing.
thorny jay
6.0.1 doesn't support magtag, so that needs to be rephrased
Can you send a signal inside?
trim elm
@thorny jay thanks for letting me know. I’m currently running an errand but I’ll fix it when I’m back
ornate breach
if you try importing in the repl you might get the same error unmangled. this i...
Oh man I missed that bit thanks
prime cove
@tulip sleet Should I close that PR and remake it so it is on a branch? I'll make the other suggested changes as well.
tulip sleet
@mild fog that's fine, sure. You can clean up your main branch by using git reset --hard to back out those commits (after using git cherry-pick or manually getting the changes) and then git push -f.
ornate breach
okay @slender iron thanks to @strange pike I got it fixed and it passes
manic glacierBOT
mild fog
<@509938944373293056> that's fine, sure. You can clean up your `main` branch by ...
(wrong adam 👍 )
tulip sleet
(wrong adam 👍 )
sorry!
mild fog
no worries at all just didn't want the other adam to miss the message
prime cove
@tulip sleet You mentioned rounding to the next power of 2 for the buffer, looking at the data sizes that could potentially add ~32kB of unneeded RAM. Would just raising an error to let the user make the call be better than that?
analog bridge
alarm.wake_alarm seems to be broken for light sleep... can someone confirm? I am on latest main.
import time
import alarm
print('entering light sleep')
alarm.light_sleep_until_alarms(alarm.time.TimeAlarm(monotonic_time=time.monotonic() + 2))
print(alarm.wake_alarm)
@analog bridge Could you open an issue? I know this was a deep sleep issue which i thought was fixed, but not light sleep.
analog bridge
sure... 🙂
I think this is related to missing function call for setting wake_alarm after a light sleep wake.
I might be able get it fixed with touch_alarm... if not I'll open an issue for it.
trim elm
@thorny jay I just updated the sentence you were concerned about. Thanks for pointing that out to me. I just assumed that 6.0.1 works with the magtag without actually checking so it's definitely good that was able to be fixed relatively early on.
analog bridge
@tulip sleet found it... I inserted alarm_save_wakeup_alarm() here :-
https://github.com/adafruit/circuitpython/blob/main/ports/esp32s2/common-hal/alarm/__init__.c#L119
tulip sleet
@analog bridge is your PR almost ready, or should we make that a separate PR? We are trying to do beta.3 very soon?
analog bridge
I am almost done.... organizing commits... will push in a few minutes.
tulip sleet
@slender iron note x:39pm above
slender iron
👍
onyx hinge
@tulip sleet if #3905 finishes please go ahead and merge it. I'll check back about the status of this sleep fix after I get home (on an errand now)
crimson ferry
do others see a lot like this in the esp32-s2 debug console when GET-ing:E (19314804) esp-tls-mbedtls: write error :-80: E (19314804) esp-tls-mbedtls: write error :-78:The GET calls are successful, and result in 200 status and full content. maybe it's just normal low-level retries?
analog bridge
I pushed touch-alarm changes along with light-sleep fix...
thorny jay
<@!681528999520108564> I just updated the sentence you were concerned about. Tha...
Great, I did not test the learn guide. I just quickly started to read and found that. Maybe I should just follow it completely and give it a try.
idle owl
@crimson ferry Are you around?
crimson ferry
yes
idle owl
Excellent. I wanted to give you a quick few sentence primer on what you can do as a reviewer since we may not have made it clear going in. 🙂
You are welcome to review, request changes on, and merge PRs. It's best to merge once you've tested it, but obviously for some, it's pretty obvious whether it's going to work or not, and at that point you can say "Approved, did not test" or some such so it's clear, and merge the PR. Once merged, please feel free to do a release. There's standard copy in all the releases about installing the bundle from circuitpython.org, and using pip, and regarding RTD. Include that always, and if you want to, you can include a note above that boilerplate saying what changed. That's typically what's in the title of the release though, so it's not always necessary. With releasing, regarding versioning, small bug fixes are x.x.N bumps, larger but not breaking changes are x.N.0 changes, and breaking changes and huge feature adds are N.0.0 changes. (Let me know if that doesn't make sense.)
I usually edit an old release and copy and paste the markdown from it to the new release and make any necessary changes to notes.
Version numbers are free for the most part, so don't fret over what one to bump, pick one and go with it.
@crimson ferry Any other questions I didn't cover?
crimson ferry
No, that's perfect, thank you!
idle owl
You're entirely welcome! Apologies for not making that clear going in. You're not the only one who was unclear. I will make sure that we introduce folks to reviewing much more clearly moving forward.
onyx hinge
whew, home again.
idle owl
Welcome back.
onyx hinge
thanks
onyx hinge
@slender iron want me to continue with the beta release today without the TouchWake changes in #3850? Or do it tomorrow (I had not planned to work tomorrow but it's no problem to put in a couple of hours and doesn't disrupt any of my plans)
tulip sleet
the fix in that PR for light sleep wake alarm is probably a good thing
even by itself
and we could merge touch with only light testing, since it's beta
onyx hinge
er that must not be the PR I mean. Let me scroll back
that one failed CI and has requested changes and is not new. I thought I saw something new go by from @analog bridge while I wasn't able to give it full attention.
onyx hinge
`alarm.wake_alarm` seems to be broken for light sleep... can someone confirm? I ...
This is the conversation/problem I was thinking of, which is not a PR yet?
onyx hinge
OK, the failed build is a red herring (upload artifacts failed)
let me see if it looks like scott's review items were addressed
manic glacierBOT
onyx hinge
I think it addresses scott's review notes..
Does the interrupt work during light sleep?
I forgot to test this with mechanics of pretend-to-sleep... wake on interrupt from light sleep is working now.
tulip sleet
@onyx hinge do you know about repos/circuitpython/tools/convert_release_notes.py, which converts the release note markdown to HTML and BBCode? Scott keeps the quote delimeters it adds, I remove them.
onyx hinge
yes except .. it's not working for me
Requirement already up-to-date: mistune in /home/jepler/.local/lib/python3.7/site-packages (0.8.4)
jepler@babs:~/src/circuitpython$ python3 tools/convert_release_notes.py < 610b3.txt
Traceback (most recent call last):
File "tools/convert_release_notes.py", line 7, in <module>
import mistune.renderers
ModuleNotFoundError: No module named 'mistune.renderers'; 'mistune' is not a package
also you can use the Duplicate Post button in the blog to duplicate a previous beta announcement. I forgot to change the permalink last time, don't repeat my error
i am using mistune 2.0.0a4
onyx hinge
ok that worked
tulip sleet
maybe I upgraded something else. checking... [no, apparently]
orchid basinBOT
onyx hinge
"new languages" -- must just be a bug?
I want to merge that PR before doing the forum and blog announces right?
- "downloads": 144,
+ "downloads": 0,``` hm what do I do about download counts !?I can update the download counts. I have the data from just a couple of days ago.
onyx hinge
OK, would you do it by adding to that PR, or as a separate PR?
tulip sleet
I update the PR
onyx hinge
Would you please?
tulip sleet
sure, i am eating, will try to do it during the meeting
idle owl
Indeed
tulip sleet
@onyx hinge ok pushed downloads update
idle owl
@ionic elk I asked Noe to post a video of their ESP32-S2 LED sign flickering because I think their might always be white. He'll add a video clip to the GitHub issue. If needed next week, I can set up my project again and get video of it flickering because I swear it wasn't only white, but maybe I'm wrong.
onyx hinge
@lone axle I love your suggestion of having the full icon list, but it uses a LOT of ram! ```>>> gc.collect(); gc.mem_free()
219296
import forkawesome
gc.collect(); gc.mem_free()
188224
hoary moat
Are there any obvious blockers to porting the WSGI web server to Feather M0/M4 plus Ethernet Wing? My end goal is a simple REST api to toggle latching relay feather wings, which will be controlling a hacked Sabrent USB 3.0 switch for automated Yocto Project testing.
ionic elk
@idle owl Ok I'll plan on spending some time recreating issues this week
crimson ferry
@hoary moat I might be oversimplifying, but it may just be a matter of replacing the esp socket with the wiznet5k socket https://github.com/adafruit/Adafruit_CircuitPython_ESP32SPI/blob/fce466bd2bb70ca86b79e5cb36bbaca00afacfd1/adafruit_esp32spi/adafruit_esp32spi_wsgiserver.py#L35 Note that wiznet5k doesn't have HTTPS currently. Also, I'd be a little surprised if M0 boards have enough RAM for this, M4 should be fine.
hoary moat
<@!746490342253133916> I might be oversimplifying, but it may just be a matter o...
Yes. That’s exactly what thought and meant. I couldn’t find the link on my phone lol. SSL/TLS is not required, it’s on a jailed subnet.
Good to know about M0 RAM. I’ll save that for other fun projects and just port the M4 (and contribute the code when it works)
Of course I have to actually order the Ethernet Feather because I only found out it exists a week after receiving the original Feather order. #2020
I realize Arduino support is out there. CircuitPython all the way.
crimson ferry
If you need more interfaces, they plug-and-play pretty well 😉 https://twitter.com/anecdat/status/1327472474624323585
hoary moat
@crimson ferry I’m contemplating maybe 4x latching relay wings... should theoretically be scalable to 10 with 20 GPIO pins...
crimson ferry
that's a huge feather sandwich 🙂
hoary moat
Now that I said it, of course I have to build the 10x version. Lol
And then write a labgrid driver https://github.com/labgrid-project/labgrid/tree/master/labgrid/driver
I see a future show and tell
paper hedge
Hi all, I'm trying to hook up a sun sensor device over I2C to a custom pull up board. There are intermittently ERRNO 19, ERRNO 5, and "SDA / SCL needs a pull up" errors when trying to use a certain set of pins. This is a long shot, but it looks like the Samd51 chip we're using has internal pull up resistors. Anyone know how to disable these in CP?
idle wharf
For circup: make check runs pyflakes and pycodestyle, CI runs Pylint I'd like to make it consistent so you get the same results locally. What is the preferred set of code checks for CP supporting tools such as circup ?
strange pike
weirdness: 6.1.0-beta.3 qtpy builds lost about ~40% of the free space that 6.0.1 builds had (w/ gcc10): only 2764b free in beta3 vs 4612b free in 6.0.1.
this makes it more difficult in my use case, but i dunno if it's concerning in general.
atomic summit
I totally missed it, but congrats to all on 6.1 beta3 ! I hope you all have a safe and happy new year!
calm rain
trim elm
Great, I did not test the learn guide. I just quickly started to read and found ...
Yeah, that would be great
supple hawk
@slender iron
Traceback (most recent call last):
File "code.py", line 42, in <module>
File "adafruit_ble/__init__.py", line 340, in tx_power
NotImplementedError: setting tx_power not yet implemented
Safe to assume there's a bit of scaffolding that'd go into enforcing correct power limits on each platform, and also the register-level code to actually set this on a given chip? Or is it actually kind of trivial and just hasn't been done?
onyx hinge
I've been working on a web-based converter for bdf -> pcf conversion. It's ready for someone to kick the tires. It works entirely in the browser and is actually hosted by github pages: http://www.unpythonic.net/web-bdftopcf/ [source code: https://github.com/jepler/web-bdftopcf]
I've learned a lot but still not how to make web pages pretty 🙂
slow matrix
I need a bit of advice since I
am new to Circuit Python. I'm trying to use an Adafruit msa301 accelerometer with a Feather M0. From what I've been able to glean so far, I am likely running into a lack of sufficient RAM issue but I want to check before I reinvest in an M4 feather. The program is:
import time
import board
import busio
import adafruit_msa301
i2c = busio.I2C(board.SCL, board.SDA)
msa = adafruit_msa301.MSA301(i2c)
while True:
print("%f %f %f" % msa.acceleration)
time.sleep(0.5)
And upon loading I get the following error message:
Traceback (most recent call last):
File "code.py", line 4, in <module>
File "adafruit_msa301.py", line 207, in <module>
File "adafruit_msa301.py", line 233, in MSA301
File "adafruit_register/i2c_bits.py", line 61, in init
OverflowError: small int overflow
This isn't an explicit out-of-memory message but it seems unlikely that there is a overflow issue built into the accelerometer library either. Can someone help me understand if I'm just low on RAM or low on understanding.
Thanks
solar whale
That is an odd error. Just to check. Are you loading the .py or .mpy versions of all the libraries? -- you must use the .mpy versions. I'm searching to see if I have one of those sensors handy...
slow matrix
Let me check on the library...
I am, indeed, loading the .mpy version from the adafruit-circuitpython-bundle-6.x-mpy-20201223.zip archive.
solar whale
OK -- just found mine -- now to find an M0 express and bring it up to date. I'll be back in a few minutes....
slow matrix
Thanks for the help, I'll get another cup of coffee and await your findings.
onyx hinge
Line 233: _xyz_raw = ROBits(48, _MSA301_REG_OUT_X_L, 0, 6)
This specifies that the "xyz raw" register has 48 bits, which it will try to treat as a single integer value. This doesn't work in M0 builds, because they don't support "long integers", they only support integers 30 bits long (31 bits? anyway, fewer than 48 and even fewer than the customary 32).
I don't know that there's a simple solution to this problem, but if you file an Issue about it the maintainer will be able to decide whether to fix it or to add a notice that the library is not compatible with M0 boards.
solar whale
That is too bad -- I'm still setting mine up, but expect to reproduce your results. I'd get the m4....
onyx hinge
I don't want to put words in anyone's mouth, but I think there's a reasonable possibility it can be modified to get rid of this particular use of a 48 bit quantity. Instead of using RWBits for this particular data, you would use a 6-byte readinto() followed by a struct.unpack. It's all a conjecture from me though
solar whale
interesting -- the msa301_simpletest from the bundle works for me ```
Adafruit CircuitPython 6.1.0-beta.3 on 2020-12-31; Adafruit Feather M0 Express with samd21g18
import msa301_simpletest
6.526161 -6.214935 4.036356
6.574041 -6.214935 3.945382
6.540525 -6.243664 3.926230
6.574041 -6.238874 3.907078
oh yeah it's a basic vs express distinction. feather m0 express has long, feather m0 basic doesn't.
solar whale
as does your code @slow matrix ```Adafruit CircuitPython 6.1.0-beta.3 on 2020-12-31; Adafruit Feather M0 Express with samd21g18
import msa_test
6.454338 -6.229300 4.170422
6.454338 -6.229300 4.074661
6.459127 -6.243664 4.098600
6.473492 -6.224510 4.093813
6.473492 -6.243664 4.089025
6.478279 -6.277180 3.998052
I am admittedly running the latest version of the feather_m0_build ---- What version are you running?
Ah I assumed you had an M0 express -- which board are you using?
slow matrix
I just got the M0 Basic a week or two ago so I suspect the latest version. How do I get the version from the device?
solar whale
The basic is a different beast than an express.
strange pike
it's printed at the top of the repl
slow matrix
Yeah, I just saw that. It is: Adafruit CircuitPython 6.0.0 on 2020-11-16; Adafruit Feather M0 Basic with samd21g18
strange pike
also in the boot_out.txt file on the flash
onyx hinge
the version is not the problem, that's very close to the current stable version
solar whale
much more limited. I'll try it on a basic, but as @onyx hinge noted it will not work....
onyx hinge
I'm pretty sure the m0 basic is going to be incompatible with that library as it stands
solar whale
confirmed -- the adalogger is a "basic" ```
Adafruit CircuitPython 6.1.0-beta.3 on 2020-12-31; Adafruit Feather M0 Adalogger with samd21g18
import msa_test
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "msa_test.py", line 4, in <module>
File "adafruit_msa301.py", line 207, in <module>
File "adafruit_msa301.py", line 233, in MSA301
File "adafruit_register/i2c_bits.py", line 61, in init
OverflowError: small int overflow
In general, the M0 basic will be a frustrating place to work with CircuitPython .... but it it worth opening an issue. Are you comfortable doing that or would you like me to open it?
slow matrix
My application will also require a 14-segment display and the native touch input so I'm pretty sure I'll need more memory than what the Basic can provide. I'm always looking for the lowest cost, leanest implementation but M0 Basic doesn't seem to be the answer here. Thanks for all your help, it is very much appreciated.
I would be happy to learn how to submit an issue as one never knows when the knowledge will come in handy. Just a pointer to where to start would help and I'll stumble around from there.
Thanks again.
solar whale
go here https://github.com/adafruit/Adafruit_CircuitPython_MSA301/issues and click on "new issue" -- include your code example -- or just reference the msa301_simpletest.py example and include a copy of the error traceback.
slow matrix
Excellent! I'll do that. You all have been very helpful.
solar whale
Make it clear you are using an M0 Basic.
You're welcome -- glad to try to help. You found a real issue!
Thanks @onyx hinge for the quick "root" cause.
@slow matrix for low cost, you might want to consider the itsy Bitsy M4 express https://www.adafruit.com/product/3800
with the display and a sensor, I would not recommend an M0 based board.
unless you need the battery charger on the feather.
slow matrix
Thanks for the recommendation, I'll definitely look into it. For this application, having the onboard battery charger is much less important than small form factor.
As for the bug report, I have a github account that has been fallow for a very long minute (like a four year long minute) and, as it seems silly to create a second account merely to report the bug, I'm in the process of resurrecting this account. It's been a day of discovery for me.
idle wharf
I'm messing around with a new stemma device (the best!) and the sample code looks like this:
import board
from adafruit_lc709203f import LC709023F
sensor = LC709023F(board.I2C())
That doesn't work on the MetroS2 or FeatherS2, you need to first create an i2c object and pass it in. Not a big deal for me, but I'm curious, do other boards have a default i2c like in that sample?
I'm wondering if this is something that needs to be updated in the samples or if it is something that needs to be implemented in the esp32-s2 port...
jaunty juniper
the feather S2 port used to be missing board.I2C but it's there now
is that what you are talking about or ?
slender iron
<@!252717193496756235>
```
Traceback (most recent call last):
File "code.py",...
I don't expect it to be too much work.
idle wharf
the feather S2 port used to be missing board.I2C but it's there now
is that what...
I'll try a newer build, thanks !
idle wharf
the feather S2 port used to be missing board.I2C but it's there now
is that what...
Hmm... I built a clean build off CP main and I still see the not implemented.
sensor = LC709023F(board.I2C())
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
NotImplementedError: No default I2C bus
I think the default is the critical word there
jaunty juniper
I think the default is the critical word there
the metro ? it definitely works on the feather, I use i2c = board.I2C() (PR3671, 12 november) maybe it needs to be added to the metro too
idle wharf
i'm using a feather s2
I'm using a PRE-RELEASE FeatherS2... (green) that might be it...
jaunty juniper
oh right it's treated as a different board
idle wharf
Thanks @jaunty juniper... that was it. board.I2C() works now on my build.
I see some other changes in the release vs. pre-release I'll ask UM about on his server. HNY !
manic glacierBOT
manic glacierBOT
supple hawk
I don't expect it to be too much work.
I've just looked through shared-bindings adapter.c and (nrf)common-hal adapter.c and.... Yeah I don't know what I'm doing 🙂
manic glacierBOT
thorny jay
I found code in the Learn Guide repo but I did not find the Learn Guide: https://github.com/adafruit/Adafruit_Learning_System_Guides/tree/master/ESP32_S2_WiFi_Tests (did I fail in my search? is there a guide on the FeatherS2 in preparation?) I believe this would have been useful for me as I re-did this based on MagTag example and other guide to test my FeatherS2. It would also be usefull for Metro S2 board: https://learn.adafruit.com/adafruit-metro-esp32-s2 cc: @atomic summit
onyx hinge
ornate breach
Curious if you expand Ram on the m0 using i2c SRAM
Probably not directly with circuitpython without some serious reworking I bet
Or maybe not at all
Not sure adding 64k of SRAM would do much but 🤷♂️
tulip sleet
Curious if you expand Ram on the m0 using i2c SRAM
Nope, sorry. Those chips don't support adding external RAM that is directly visible to the CPU.
ornate breach
I figured, but was worth asking 🙂
lone sandalBOT
ornate breach
@solar whale I think the issue might be in the bus_device library
I'll investigate further
it doesn't throw without a call for adafruit_bus_device
solar whale
Interesting.. Good luck finding something. There has been a lot of work done recently to move bus_device into the core. It was implemented then reverted due to some issues that are being resolved. https://github.com/adafruit/circuitpython/pull/3612
ornate breach
It's related to this specifically
https://github.com/adafruit/Adafruit_CircuitPython_BusDevice/blob/eb4b21e216efd8ec0c4862a938e81b56be961724/adafruit_bus_device/i2c_device.py#L102
tulip sleet
@ornate breach are you testing with beta.2? We turned off the native adafruit_bus_device in beta.3 because of a number of issues.
ornate breach
no
current version for my CP Sapling: Adafruit CircuitPython 6.1.0-beta.1-8-g8301dcada-dirty on 2020-11-23; CP Sapling M0 with samd21e18
local build, but the qt py is using CP 6 from 11/13
stable
both throw the same OS Error
manic glacierBOT
ornate breach
I wonder if maybe the i2c buffer isn't cleared properly on reset
I may add a delay between my write bytes section of reset() to see if that fixes it
nevermind... there is a 1ms delay in the reset write calls.
hearty tapir
@ornate breach, I think that i2c no longer has any „deinit“ (uninstall) - which was the way to fix the reload issue
hearty tapir
Maybe, but the uninstall was removed from the i2c code
hearty tapir
I‘ve spent a lot of time on this and couldn’t figure out how to fix it until Bruce removed the uninstall from the deinit function
ornate breach
right, but the issue seems to still exist at least in some builds
at least for m0
it doesn't seem to just affect i2c either
it was happening with i2s for someone in the CP help channel
hearty tapir
I have to admit, that I don’t like the solution to remove the uninstall. It is a workaround, but I hope that we can find a better way go fix it sometimes soon
ornate breach
yeah, I've honestly look over all the i2c related files, and I'm not sure where it might be even hanging up.
in my issue, mine returned for python self._i2c.write(...)
hearty tapir
I‘ve read a comment where the esp32 was still waiting for a reply from the i2c device (e.g. during discovery) and that caused some serious issue
ornate breach
I just wonder if the buffer isn't clearing
because it is also hanging on python self._write_then_readinto(...)
though, I probably could just use python self._readinto(...)
but I'm not confident that you could avoid that
hearty tapir
I haven’t run into issues again, but I mainly spent time on enhancements
rustic hornet
My NYE state machine https://github.com/bwolfram1/NYECountdown/blob/main/main.py
ornate breach
it's mostly random when it throws the OSError
it could throw it right away, or 5 minutes after running just fine.
though... the adafruit_bus_device does say that the OSError it throws is from the device not being present
I could try with other i2c devices, but other people have similar i2c issues so it might not be trivial
atomic summit
@thorny jay I don't think a per board guide is needed. Not for CP functionality, as ESP32-S2 is ESP32-S2 - There is only the 1 MCU right now, just being delivered in various formats. So from a code point of view, all of the non hardware specific stuff like Wifi is the same.
Unless there was something else FeatherS2 specific you though needed to be made clearer?
thorny jay
Well, 10 minute ago I figured there is a light sensor... I dunno if there is a demo code for it.
Also I tested TouchIO (actually I want to test wake on touch) and I took pin 33 at random, but later I looked at the schematic and now I know that I have to choose a pin that is touch compatible.
The point is/was that there is code in the Learn Guide repo, and I don't know what guide link to it.
I had no trouble to make the wifi work, it work I guess like for other board since I succeeded. 🙂
jaunty juniper
the light sensor could bear better documentation, and a place in the demo file too (light = analogio.AnalogIn(board.AMB))
atomic summit
I have more example code I am preparing that will ship with the featherS2 in the future and I'll also have on my feathers2.io website, that includes a wifi example, LDO2 use and I can also add the lightsensor.
thorny jay
Maybe like on the MagTag that also has alight sensor that is used in this guide: https://learn.adafruit.com/twin-peaks-light-reactive-pyportal-picture-frame
atomic summit
Touch is the same on all boards. it's hardware level which IO it's on, and I have those on the pinout guide.
thorny jay
Yeah, I am on that image too.
hearty tapir
Touch is the same on all boards. it's hardware level which IO it's on, and I hav...
How did you create this graphic? I really like it 🙂
atomic summit
I should add the RTCGPIO pins as well. It's on my list of things to do... I'm just have 2 hands, so struggle to cover everything I need to get done 😦 #growingpains
thorny jay
The LDO2 I tried to figure out how to use it... I guess I have to power external device from that pin, and then they will sleep when I sleep the ESP32S2.
atomic summit
I made it by hand in Affinity Designer (like illustrator)
hearty tapir
Nice, I even have that on my MBP 🙂
atomic summit
You can turn LDO2 on and off with IO21, but if the ESP32S2 deep sleeps, it auto shuts down for you.
thorny jay
How did you create this graphic? I really like it 🙂
Oh and I want that graphic in higher resolution, just because my eyes are not that good anymore... and did not go to they eye doctor yet.
hearty tapir
It is a vector graphic, so I guess @atomic summit could scale it as he wants 🙂
thorny jay
Adafruit has a link in learn guide that contain multiple resolution, I always go to the bigger one.
atomic summit
Is this ok? I can make it any size 🙂
thorny jay
Thanks, saved. Now I can print a poster and put that in my room.... not sure my wife will like that. 🙂
jaunty juniper
adafruit has pinout graphics, they might have a way to generate them ? ask an engineer 😉
thorny jay
It's not clear how they do. Depend on the guide and who does it. I have catch a conversation here where kattni needed to update one.
atomic summit
I'm happy making my pinout cards the way I do. I mean, it takes a lot of time and effort, but the I'm happy with the results and consistency across all of my products.
jaunty juniper
they look really nice
manic glacierBOT
slender iron
slender iron
this is neat: https://github.com/johnmcfarlane/unformat
ornate breach
this is neat: https://github.com/johnmcfarlane/unformat
Same person who did the clang format scripts?
slender iron
¯_(ツ)_/¯
timber mango
Filename misnamed (?) in MagTag Project Selector ... adafruit_matrixportal lib has network.mpy file ... Call from project selector weather.py is looking for network.py ... I'm just gettin' started with this CPY stuff so I'm not sure how 2 fix the coding ... Using 20210101 libs & uf2 6.1.0-beta3 ... Where do I go next ?
timber mango
Hola ... Figured it out (Note 2 self - don't work on MatrixPortal Weather B4 workin' on MagTag Weather) ... Copied wrong weather.py 2 MagTag Project ... RTM a couple of times & triple check
what U just did
calm rain
// TODO: Talk openly, develop openly