#circuitpython-dev

1 messages · Page 63 of 1

lone axle
manic glacierBOT
#
bablokb
[adafruit/circuitpython] Issue opened: #9085 Pico-W memory bloat

CircuitPython version

Adafruit CircuitPython 9.0.0 on 2024-03-19; Raspberry Pi Pico W with rp2040

Code/REPL

import gc
print(f"free memory at start: {gc.mem_free()}")

main.py Ausgabe:
free memory at start: 126112

Behavior

With CP 9.0.0, we have about 123K of free memory. With 8.0.5, we had 152K (139K with 8.1.0/8.2.10). So we lost 20% of available memory from one major release to the next.

This really hurts. Using e.g. a HTTPS-API and up...

slender iron
#

The best way to debug this is with a logic analyzer and comparing what the arduino example does vs CP's output.

fringe yarrow
# slender iron The best way to debug this is with a logic analyzer and comparing what the ardui...

I attemtped that, but the rp2040 as a logic analyzer doesn't seem to be able to keep up with the speed of the data, even when I set the parallelbus frequency down to 4mhz, the lowest it will go and still turn the screen on. The results it got were inaccurate and spotty, with some of the init sequence being visible in the decoded data stream, but a lot of it not matching what is defined in code, becuase of missed bits

slender iron
#

huh, that's surprising

#

you could try capturing just the clock line first

#

to see if the parallel version is what slows it down

fringe yarrow
#

which is the clock line?

slender iron
#

actually, parallel doesn't call it clock I think

#

maybe "write"?

#

it's been a little while since I did it

fringe yarrow
#

Has anyone confirmed recently that it works on the tdisplay?

slender iron
#

I can

#

but I haven't heard anything

manic glacierBOT
tidal kiln
#

@tulip sleet @gilded cradle still seeing the qualia wifi weirdness using 9.0.0 with libs from 2024/03/23 bundle

Adafruit CircuitPython 9.0.0 on 2024-03-19; Adafruit-Qualia-S3-RGB666 with ESP32S3
>>> from adafruit_qualia import Qualia
>>> from adafruit_qualia.graphics import Displays
>>> qualia = Qualia(Displays.ROUND28)
>>> qualia.get_local_time()
Connecting to AP peanuthamper
Getting time from IP address
Traceback (most recent call last):
  File "adafruit_connection_manager.py", line 271, in get_socket
OSError: [Errno 116] ETIMEDOUT

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "adafruit_portalbase/__init__.py", line 492, in get_local_time
  File "adafruit_portalbase/network.py", line 270, in get_local_time
  File "adafruit_portalbase/network.py", line 237, in get_strftime
  File "adafruit_requests.py", line 591, in get
  File "adafruit_requests.py", line 525, in request
  File "adafruit_connection_manager.py", line 282, in get_socket
RuntimeError: Error connecting socket: [Errno 116] ETIMEDOUT
>>>
tiny peak
slender iron
tiny peak
slender iron
#

it can also fit more stuff 🙂

fringe yarrow
#

@slender iron so I just hooked that same parallel ili9341 external screen to an ESP32-S3 and it works.. so.. i guess the issue is with the OG ESP32

gilded cradle
slender iron
#

you could try my idf 5.2 PR.

#

maybe espressif fixed it in the idf

manic glacierBOT
#
bill88t
[adafruit/circuitpython] New comment on issue #9085: Pico-W memory bloat

The primary ram hog is the wifi stack really. I don't think much (ram) can be gained from custom builds.
The difference between CP versions is probably just from the underlying lib version changes.

Some tips:

  • Consider switching to / rewriting smaller more task focused modules.
  • Consider using internal modules (like displayio for the display).
  • Consider using mpy-cross to partially-compile .py files.

If these are not enough for your applications, you may wanna look into ESP3...

fringe yarrow
slender iron
#

you can see my branch name there

onyx hinge
midnight ember
#

Those are called lithography prints where details come through via backlight.

#

I've never seen one that rotates with a servo, very cool.

turbid radish
#

👋

slender iron
fringe yarrow
#

```/home/circuitpython/.espressif/tools/xtensa-esp-elf/esp-13.2.0_20230928/xtensa-esp-elf/bin/../lib/gcc/xtensa-esp-elf/13.2.0/../../../../xtensa-esp-elf/bin/ld: build-sunton_esp32_2432S022c/esp-idf/esp-idf/wpa_supplicant/libwpa_supplicant.a(esp_wpa3.c.obj):(.literal.wpa3_build_sae_msg+0x18): undefined reference to esp_wifi_sta_get_rsnxe' /home/circuitpython/.espressif/tools/xtensa-esp-elf/esp-13.2.0_20230928/xtensa-esp-elf/bin/../lib/gcc/xtensa-esp-elf/13.2.0/../../../../xtensa-esp-elf/bin/ld: build-sunton_esp32_2432S022c/esp-idf/esp-idf/wpa_supplicant/libwpa_supplicant.a(esp_wpa3.c.obj): in function wpa3_build_sae_commit':
/home/circuitpython/repos/circuitpython/ports/espressif/esp-idf/components/wpa_supplicant/esp_supplicant/src/esp_wpa3.c:37:(.text.wpa3_build_sae_msg+0x4f): undefined reference to `esp_wifi_sta_get_rsnxe'

slender iron
#

@onyx hinge what are using to make your pycon slides?

onyx hinge
#

@slender iron google (sigh)

slender iron
#

and clean and rebuild

slender iron
manic glacierBOT
onyx hinge
#

it was Scott who added those warnings, and I'm sure he's delighted to hear it because it has been a source of concern for some. nice to know some are helped too.

slender iron
#
  • but daisy is unsupported
turbid radish
#

❤️ @idle owl

slender iron
fringe yarrow
slender iron
gilded cradle
#

Forgot to give you a hug @lone axle for also making a PR that fixes the sorting first.

slender iron
fringe yarrow
#

I suspect this ESP32 vs S3 is the answer.. since I hooked the same model screen to the same model esp32 and got the same result as that board, but that screen worked on an S3 with no other changes..

onyx hinge
#

should have let it get to 1000 days old

#

vintage bugs

fringe yarrow
#

I think I missed a step.. When I'm checked out on your idf5.2.1 branch, done the submodules, did the esp-idf/install.sh, it still looks like esp-idf 5.1 is installed

austere acorn
#

Ok, I didn't add it in the document, since this is the first major CircuitPything thing I'm working on, and just didn't think to do it... but I'm working on porting CP to the Daisy Seed board.
Figuring out how to do XiP on the 8MB external Flash is not easy, but the board stats are too tempting to not try! 😻

slender iron
slender iron
fringe yarrow
slender iron
onyx hinge
#

I think "good first issues" are specific to every repo, having typing as "good first issue" was sort of odd in our history

short tendon
#

We could create issues to validate that example code in libraries are working

#

Every library could have a .pyi file that the typing library pulls in...

onyx hinge
#

fulfulling is a good word

austere acorn
#

I agree that typing issues are **not ** interesting (ahem...boring... ahem) as first issues 😄

onyx hinge
#

+1 to opening issues. I investigated the ssd1306 driver (see more notes in the document) -- it was intended to work with 8.x and 9.0 without warnings, but a mistake was made and it needs to be revisited.

#

historically we ran pylint with different flags on examples

slender iron
onyx hinge
#

@slender iron at the moment it's not designed to be flexible, it's designed to do exactly what our demos needed. Not idea, but that's the situation.

manic glacierBOT
onyx hinge
#

No April Fools 🙂

#

thanks Tim!

short tendon
#

Thanks!

austere acorn
#

yep! 😄 @midnight ember @slender iron

#

a lot of audio stuff is not implemented for stm32

#

no, no st

#

teensy is imx

#

it's not open source anymore 😑 they closed it a few months ago

#

frustrating asf

#

there are schematics for older revs, if you dig hard enough (i have rev 5 btw... you should have the last one, rev 7, @midnight ember so we will need test both)

#

sure, thank you! 🙂

proven garnet
#

It would be a lot like the composite GitHub actions I suppose

onyx hinge
#

It looks like pre-commit-hooks.yaml can't have "- repo" blocks so you'd have to pip install in a venv or something...

proven garnet
#

Pre-commit should handle the venv part - I think we would just add ruff and the others to the requirements file of the hooks package.

#

When it installs the hook, it should handle all the dependencies as well

short tendon
#

That was my thought. Didn't know how worth the work that would be.

Also, what happens when we update the rules and someone's simple PR is now a lot bigger?

proven garnet
short tendon
#

So, a user changes 1 line, and then has a bunch of unrelated formatting changes because the global ruff file was changed

lone axle
#

Here is the notes document for next Monday’s CircuitPython Weekly Meeting. It is at the normal time of 11am Pacific / 2pm US Eastern here on Discord. Add your hug reports and status updates to the document before the meeting. If you are unable to attend but would still like to contribute, feel free to add your notes and we’ll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/1fwE8nNrxx4lDEbSlGUqE9iR_XqECxSljZI2_6P-UdDw/edit?usp=sharing

Google Docs
CircuitPython Weekly Meeting for April 1, 2024
lone axle
proven garnet
#

It's turtles the whole way down

short tendon
#

It's a hard one. I like the idea of a source of truth, but it seems more complicated than it's worth...

crimson ferry
short tendon
#

Also might be strange to see a PR with a bunch of formatting changes, but no configs changing...

manic glacierBOT
#
bablokb
[adafruit/circuitpython] New comment on issue #9085: Pico-W memory bloat

I have done already all of this (e.g. use pre-compiled modules) and have carefully garbage collected my objects. All my tests are with the exactly same code and library versions. I have instrumented every part of my program and have detailed data for various modules I use.

The data clearly shows that wifi and other module usage does not differ greatly between CP versions. The big initial difference of free-mem at the beginning of the program carries through the whole runtime of the program...

wraith crow
#

I've been turning over ideas about good first issues and I think one of the easiest ways to contribute is to add a new board, I don't think we want 300 new boards added just for the sake of adding boards through. I wonder if there would be a way to design a playground board with all sorts of gadgets just for the purpose of teaching people how to add/fix CP features. The updates could be built and tested on the board (always rewarding to see CP do something new on hardware that you had a hand in) but maybe not even merged in so they could be added again by the next students/tinkerers.... Just thinking out loud....

austere acorn
#

I think one of the easiest ways to contribute is to add a new board

Speaking as someone who is just doing a board port as the first major CP project, I don't think I agree 😆
Maybe adding another samd21 board without many extra peripherals, so the port from another board can be just copy pasted, but I still think it would be intimidating 🤷‍♀️

manic glacierBOT
#
tyeth
[adafruit/circuitpython] Issue opened: #9086 ItsyBitsy ESP32 can't use onboard NeoPixel in v9.0.0

CircuitPython version

Adafruit CircuitPython 9.0.0 on 2024-03-19; Adafruit ItsyBitsy ESP32 with ESP32

Code/REPL

# import time
# import board
# import neopixel

# pixels = neopixel.NeoPixel(board.NEOPIXEL, 1)

# while True:
#     pixels.fill(0xADAF00)
#     time.sleep(1)
#     pixels.fill(0)
#     time.sleep(1)

import board
import digitalio
import neopixel
import time

print("START")
# On MagTag, enable power to NeoPixels.
# Remove these ...
dense jasper
#

hello community,

does anybody knows of any activity / efforts to program circuitpython with a VPL / node flow based UI / roberta (scratch) like / similar?
iam interessted in experimenting with this for an even easier start for beginners...

proven garnet
short tendon
#

@proven garnet please let me know how I can help on the ruff stuff. In theory if you take the few changed files (maybe moving the config to ruff.toml per Dan) and flag any library that has issues, I can work on manually updating?

#

I can also create a branch with only the changed files...

proven garnet
short tendon
#

Would it be helpful if I updated requests and connection manager to match on the pre-commit file? Since they both have isort?

slender iron
slender iron
manic glacierBOT
lone axle
#

One idea I had about potential good first issues is sensor + display examples. Right now I think there are a lot of sensor driver libraries that have examples that only print data to serial console. Many times on the product pages there is gif or video of an example running that updates the display with values from the sensor. I assume the ones used for the video are early in the devices life and are arduino based. I do think it would be nice to have CircuitPython display examples available in those sensor libraries though.

One thing we've seen folks have trouble with a few times over the years is getting the label created and updating it without creating new ones. That is documented in a DisplayText learn guide now, but it would be nice to have relavent examples to point someone toward for any given sensor they're working on.

manic glacierBOT
#
tannewt
[adafruit/circuitpython] New comment on issue #9085: Pico-W memory bloat

The starting heap memory is largely dictated by the static memory allocations. Features like picodvi can use this space up because it includes all code that needs to run when flash is unavailable (like the dvi conversion code.)

To dig into this, you'll want to do your own build and look at the .map file. It'll lay out how much of RAM is used statically and what symbols take up that space.

#
tannewt
[adafruit/circuitpython] New comment on issue #9079: IDF reports resource not available on soft reboot

Possibly related to #9086 but unlikely as probably user error and a reasonable enough error message. Closing

This is probably an issue. A simpler example like the one @RetiredWizard is how to know for sure.

It turns out that at least on the QT PY, the the AnalogIn object for board.A2 doesn't have a "value" function:

I think this is a different bug around how dir() figures out what is on an object.

orchid basinBOT
manic glacierBOT
orchid basinBOT
manic glacierBOT
#
mrklingon
[adafruit/circuitpython] Issue opened: #9088 Memory Allocation Fail with Neo Trinkey

CircuitPython version

Memory problem with CP 9.0.0 - code fails that works with CP 8.2.10

code.py output:
]0;🐍code.py | 9.0.0\Traceback (most recent call last):
File "code.py", line 1, in 
MemoryError: memory allocation failed, allocating 1016 bytes
]0;🐍1@code.py MemoryError | 9.0.0\
Code done running.

-------


Auto-reload is on. Simply save files over USB to run them or enter REPL to disable.
code.py output:
]0;🐍code.py | 8.2.10\TREK
-  .-.  .  -.-  

Code:...
#
bill88t
[adafruit/circuitpython] New comment on issue #9085: Pico-W memory bloat

One more tiny tip I have is that you may wanna consider setting CIRCUITPY_PYSTACK_SIZE to something like 1792 or even lower.
For every 256 you remove from this value it, you should gain 256 bytes ram but lose a few stack frames.
This means lower function depth.
If you have a lot of nested calls (or your libs do), you will see pystack errors with this change though.
You can fine tune it and have exactly the size you need though.

fringe yarrow
#

@slender iron sorry to keep on about this.. I'm not sure if I managed to fetch your 5.2.1 PR or not.. I think i did.. if so, it didn't make any difference.. the OG ESP32 still doesn't talk to the parallel LCD

slender iron
#

the debug output will tell you want idf version it was built with

fringe yarrow
#

which would be a simple thing if this didn't have s uch little flash 😄 enabling debug means remembering the arcane list of things i had to disable again

manic glacierBOT
manic glacierBOT
#
tyeth
[adafruit/circuitpython] Issue opened: #9089 Can't vary frequency of PWM pin on ItsyBitsy ESP32 - assignment ignored

CircuitPython version

9.0.0 itsybitsy esp32

Code/REPL

# SPDX-FileCopyrightText: 2024 Tyeth Gundry
# SPDX-License-Identifier: MIT

#TODO: Add Adafruit IO feeds for BME688.
#TODO: Add scaled output to air quality feed from scaled air sensor (CO2/PM2.5/gas-ohms)
#TODO: swap to using rotary switch inputs ( max_events=1 ) with keypad.keys:
# https://docs.circuitpython.org/en/latest/shared-bindings/keypad/index.html#keypad.Keys

#NO LONGER TRUE, WAS FOR Q...
fringe yarrow
#

@slender iron ESP-IDF v5.1-dev-9318-g722fba96a1 clearly I dont know enough about git to know how to get whats needed to compile with 5.2.1

slender iron
manic glacierBOT
#
tyeth
[adafruit/circuitpython] New comment on issue #9089: Can't vary frequency of PWM pin on ItsyBitsy ESP32 - assignment ignored

strangely my other longer code is working now having switched the base frequency from 100 to 300, the auto-adjusting code is now kicking in as expected and my frequency is varying. Not retested the test case, but it's in the same code.py just commented out, no power off of the device has happened yet.

I'm now turning off the itsy (its in a device). I'll look properly tomorrow evening

fringe yarrow
#

I pulled a fresh clone from adafruit/circuitpython, then used github cli and gh pr checkout 9071

brazen hatch
#

I also have trouble with idf5.2.1..
I checked out the branch (I did fetch all right before)

[bill88t@lunchbox | ~/git/circuitpython/ports/espressif]> gs
On branch idf5.2.1
Your branch is up to date with 'tannewt/idf5.2.1'.

nothing to commit, working tree clean
#

but I am getting

-- Checking Python dependencies...
ERROR: /home/bill88t/.espressif/espidf.constraints.v5.1.txt doesn't exist. Perhaps you've forgotten to run the install scripts. Please check the installation guide for more information.
CMake Error at esp-idf/tools/cmake/build.cmake:365 (message):
  Some Python dependencies must be installed.  Check above message for
  details.
Call Stack (most recent call first):
  esp-idf/tools/cmake/build.cmake:502 (__build_check_python)
  esp-idf/tools/cmake/project.cmake:605 (idf_build_process)
  CMakeLists.txt:12 (project)

during the compilation, AFTER the idf has been loaded already

# 
Setting IDF_PATH to '/home/bill88t/git/esp-idf-v5.2.1'
Detecting the Python interpreter
Checking "python3" ...
Python 3.11.8
"python3" has been detected
Checking Python compatibility
Checking other ESP-IDF version.
Adding ESP-IDF tools to PATH...
Checking if Python packages are up to date...
Constraint file: /home/bill88t/.espressif/espidf.constraints.v5.2.txt
Requirement files:
 - /home/bill88t/git/esp-idf-v5.2.1/tools/requirements/requirements.core.txt
Python being checked: /home/bill88t/.espressif/python_env/idf5.2_py3.11_env/bin/python
Python requirements are satisfied.
Added the following directories to PATH:
  /home/bill88t/git/esp-idf-v5.2.1/components/espcoredump
  /home/bill88t/git/esp-idf-v5.2.1/components/partition_table
  /home/bill88t/git/esp-idf-v5.2.1/components/app_update
[...]
  /home/bill88t/git/esp-idf-v5.2.1/tools
Done! You can now compile ESP-IDF projects.
Go to the project directory and run:

  idf.py build
#

I did wipe submodules as always and even made a new CP clone

brazen hatch
#

I think the forked idf has something not playing nice with my external stock idf.
But I have been always having a seperate one. Why did it break now?

manic glacierBOT
main furnace
#

Seems to be a Windows 11 problem. Tried two different picos, same result. Code.py is just print("Hello world!").

Works fine on Fedora 29.

image.png
manic glacierBOT
#
dhalbert
[adafruit/circuitpython] New comment on issue #9085: Pico-W memory bloat

I did some looking at this, first by just using arm-none-eabi-nm, and then by using https://github.com/google/bloaty (which I had never used before)

Here is some info, first comparing 8.1.0 to 8.0.5. Notice the growth in the in the .data and .itcm sections, which are RAM. .itcm is code in RAM (right?). .rodata can be flash.

I need to build without picodvi turned on, which I didn't do consistently yet.

$ ~/repos/bloaty/build/bloaty ~/build-raspberry_pi_pico_w_8.1.0/firmwa...
manic glacierBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
#
xs5871
[adafruit/circuitpython] New review comment on pull request #8855: Add state integration based debouncing to the `keypad` module

I did test with longer intervals, works as expected, but then I'd have to set the integration threshold lower to get reasonable response times (obviously). I intentionally tested the extreme case of lowest latency -- increasing any of the two variables will always increase both reliability and latency.

If you'd like me to contribute to that write-up, maybe with a graph illustrating the algorithm as you mentioned, I'd be happy to help.

dense jasper
orchid basinBOT
manic glacierBOT
midnight ember
#

We'll work it out. As long as Scott says it's doable then it'll just take time and a lot of research.

manic glacierBOT
#
tyeth
[adafruit/circuitpython] New comment on issue #9089: Can't vary frequency of PWM pin on ItsyBitsy ESP32 - assignment ignored

Just to add it's still set at 300Hz and switched on fine this morning.

Leads me to believe it may have been from initialising the pwm pin at a fixed frequency of 100Hz when testing code initially (no variable frequency or False effectively), then soft rebooting and initialising at 100Hz with variable frequency True (no deinit performed?), at which point variable frequency was ignored.

manic glacierBOT
#
anecdata
[adafruit/circuitpython] New comment on pull request #9084: raspberrypi: implement setsockopt(SOL_SOCKET, SO_REUSEADDR)

With today's artifact and same latest release of the adafruit_httpserver library, I get OSError: [Errno 112] EADDRINUSE upon bind by default when the server is interrupted (^C) and restarted (^D). This code is in server.py:

        if implementation.version >= (9,) or implementation.name != "circuitpython":
            sock.setsockopt(socket_source.SOL_SOCKET, socket_source.SO_REUSEADDR, 1)

But if I change it to `sock.setsockopt(socket_source.SOL_SOCKET, socket_source.SO_...

manic glacierBOT
manic glacierBOT
proven garnet
#

@tulip sleet @short tendon Catching up on the meeting from yesterday, I think a centralized ruff.toml could be interesting. If we create our own pre-commit hook we could do a lot. I think the "being offline" issue isn't a problem because eventually they'll need to git push anyway so Internet access is an eventuality. We could use the venv pre-commit creates to run the tools we want via Python instead of command line, assuming they all have Python APIs (or use subprocess.Popen).

#

I think ruff.toml would be the only thing that would be factored out since the other infrastructure really is needed on a per repo basis. But unifying REUSE, ruff, etc. would be chefskiss

#

I do think it's feasible to do it, thinking about it. I could prototype it after the ruff patch if we want.

tulip sleet
proven garnet
#

It does mean we probably won't run ruff as a pre-commit hook

#

So that's a good point

tulip sleet
#

i think running ruff locally in pre-commit is important, just as pylint and black are now important

#

it's not terrible to have a ruff.toml or incorporate those settings into pyproject.toml. The dread is if ruff gets updated and causes a lot of code churn. Is the version of ruff that would be used by pre-commit specified explicitly, that is, does it have to be specified explicitly? pre-commit has very particular ideas about certain things

proven garnet
#

It would still run locally, but it would be our pre-commit hook running it instead, still within a venv. So our hook is installed and it has ruff and other tools as dependencies and runs those within that venv.

proven garnet
tulip sleet
#

also i don't have a strong opinion about leaving the settings in pyproject.toml vs a separate ruff.toml. It's easier to patch only the ruff settings if it's separate, I guess, that leans towards a separate ruff.toml

proven garnet
#

But let me double check that

#

I do agree that ruff.toml is easier to patch and preferable in that regard

orchid basinBOT
tulip sleet
proven garnet
#

Sorry, GitHub, not PyPI

tulip sleet
#

so the separate black and pylint entries would be moved into the hooks in lines 17-19? But those hooks are supplied by pre-commit, we might need our own. I should look and see whether pre-commit has already done anything with ruff

onyx hinge
tulip sleet
#

there is special handling for examples vs library code in that pre-commit .yaml, for instance, we need to be able to do that

proven garnet
#

Ruff has pre-commit hooks, but would need local settings in each repo for customization. If we make our own hook, we could "internally" run the tools, but with the benefit of being able to use "internal" default settings (i.e, a ruff.toml).

tulip sleet
#

and vary the running on the different categories of .py files? I guess our hook could do anything we want it to

proven garnet
#

I guess the way I think about it is that we'd effectively creating a wrapper/combo CLI that runs all the tools and make a hook for it.

proven garnet
#

It does also mean that updating the hook versions is also easier because we just change the dependency in the combo tool.

tulip sleet
#

as we've already done when factoring out other CI tasks (not pre-commit)

proven garnet
#

The more I think about it, the more of a neat solution it is, though it does mean the work @short tendon has done will need to get transferred to the new tool.

tulip sleet
#

Justin said he was just experimenting with several representative libraries. That is all draft work and I think he would say it's fine to redo it in a different form. So I think the first thing to do is to set up the hook and then PR against one or more of those libraries (maybe suggesting modifications to Justin's PR's) to test it.

proven garnet
#

Sounds good! I'll get on it!

tulip sleet
#

Thank you!!

orchid basinBOT
slender iron
brazen hatch
#

I got it from releases as usual

slender iron
#

did you run the install script?

brazen hatch
#

yea

#

it's running normally, the deps are ok

brazen hatch
#

my aliases are conplicated, but they haven't changed since idf 4.2

#

I just have been changing the version the same way up to now

#

I have always been keeping my own idf that matches cp version

slender iron
#

what error are you getting?

brazen hatch
#

it's as if it's trying to find 5.1 stuff

#

I can get a full log if you want

slender iron
#

that doesn't seem CP related

#

that looks idf related

brazen hatch
#

yea but the external idf works (I built examples)

#

at some point the makefiles are setting it idf version to 5.1 I think

#

in either case I will try in docker next

slender iron
#

the CI built just fine

#

so I don't think it's CP's bug

manic glacierBOT
slender iron
manic glacierBOT
#
tannewt
[adafruit/circuitpython] New comment on issue #9089: Can't vary frequency of PWM pin on ItsyBitsy ESP32 - assignment ignored

Just to add it's still set at 300Hz and switched on fine this morning with speed varying as expected.

Leads me to believe it may have been from initialising the pwm pin at a fixed frequency of 100Hz when testing code initially (no variable frequency or False effectively), then soft rebooting and initialising at 100Hz with variable frequency True (no deinit performed?), at which point variable frequency was ignored.

Power cycling obviously cleared the state / masked the problem...

proven garnet
manic glacierBOT
slender iron
lone axle
# slender iron FYI <@382939733107408897>

I don't think I really have a strong opinion one way or another between having compatible code that supresses the warning vs. not having compatible code that prints the warning.

But I do know that several libraries already had this change, I was (still am really) in the process of grepping the bundle to find them. The ones I've submitted so far I basically copy / pasted the changes from the last round that were completed.

At this point either way we will need to go back and do another round to switch to the 9.x only way after the 8.x bundle is finalized. Skipping these ones now would avoids one change in the interim but does leave the libraries in an inconsistent state since some already have it and some don't.

slender iron
#

ya, I realize I'm late to the party

wraith crow
proven garnet
lone axle
#

I will say I look at making the changes earlier as a bit more than just supressing the errors. I look at it also as a way to communicate the new "correct" way to do whatever the thing is. So I do kind of like it when that finds it's way into the libraries before it's absolutely necessary for them to run or else they crash without the change.

proven garnet
lone axle
#

But I also look at and copy from library code and examples quite frequently, I'm not sure if that is really true of the majority of the user base though.

slender iron
slender iron
lone axle
slender iron
#

since 9 supports both ways

#

later versions of 9 can have the warnings added back too

#

before 10 is stable

wraith crow
slender iron
#

I'd rather wait to inform them when they don't need to support the old way too

wraith crow
proven garnet
#

At the very least you can programmatically find specific code

slender iron
proven garnet
#

And manually fix them

lone axle
manic glacierBOT
lone axle
# proven garnet `libcst` can parse a concrete syntax tree as well as make changes. And since th...

I see, that does sound handy, in this case there are a few tweaks needed outside of the imports as well due to types and usages referencing things like displayio.FourWire in the ones I changed them to FourWire only and have the imports to switch which module it came from, that way a future change of this type would only require change to the import, but I don't know the likelyhood that it comes up again.

slender iron
lone axle
slender iron
#

@tulip sleet ☝️

tulip sleet
#

@lone axle The libraries all have conditional from-imports that first try the 9.x naming and if that fails, they do from-imports the 8.x way. That way avoids any warning printings. I would suggest doing the Learn Guide code that way for now, and then undo the 8.x naming in both the guides and the libraries.

#

Limor's input was just that you should do it instead of me 🙂

lone axle
tulip sleet
#

but we already did that on a bunch of the libraries

lone axle
#

I know, I copied from them for the newer ones I've submitted.

#

I can go back and update the PR in Learn to make it compatible with both, but that sounds like code that Scott prefers not to have in the Libraries and I assume Learn though it wasn't mentioned specifically. I'm willing to go through and do it whichever way is decided is best.

orchid basinBOT
tulip sleet
#

@slender iron @lone axle We did dozens of libraries or more with the conditional imports a few months ago, so I think it's OK to continue that trend for consistency. We could also turn off the warnings and just delay updating the libraries until we stop making 8.x bundles. So hold off on any new PR's now, and make 9.x-only changes after 8.x is no longer supported. we'll release a 9.0.1 soon anyway with other fixes and we can turn off the warnings then.

It's really fine with me either way. It's less work not to make these changes now and just do the 9.x-only at the right time.

slender iron
#

We can add them back when we stop supporting 8

#

in prep for 10

tulip sleet
#

we will stop making new 8.x bundles sooner than that, I think. We stopped 7.x bundles pretty early after 8.0.x. What time frame were you thinking of? Maybe 9.1.0 final?

#

last 7.x bundle is 20231003

slender iron
#

and 8.x was released in 202302

#

I thought we generally keep the last version going until we're wanting to do pre-releases for the next major version

tulip sleet
#

maybe so, I am misremembering. OK, so we can leave the current dual-import libraries alone, since they work fine

slender iron
#

ya, we can leave them for now

tulip sleet
#

do you have a PR for turning off the warnings?

slender iron
#

this is work I'd expect to do in prep for 10

#

I don't

orchid basinBOT
slender iron
#

I'm in other weeds atm

#

that's where they are generated

lone axle
#

To make sure I understand, I'm about half way through the grep results looking for instances of displayio.FourWire in libraries and example code and changing them to use the try/except importing new way and fallback to old way. Should I continue my way through that list when I can get back to it, or stop where I'm at?

And on the Learn Guide code what should be done there? Nothing further at this time? or go back thru and update with the try/except similar to the libraries?

Lastly how about Learn Guide pages, I've updated all of the relavent pages for changes from that PR already with the 9.x version of code... should I go back and revert those to 8.x or the try/catch fallback version? or just leave them in the new 9.x style?

tulip sleet
slender iron
#

The other option is that we could drop 8.x support now

slender iron
#

We shouldn't add code that has to try/except to support two versions.

tulip sleet
slender iron
#

9 supports both so that we can wait to move once we're confident in 9

#

@lone axle I think we leave all try/except stuff that's been done already. Learn guides should still use the old way because we're not stopping 8 support yet.

tulip sleet
slender iron
lone axle
#

Okay sounds good, I converted that PR to a draft so it can hang out there until we're ready for that change to 9.x syntax. I'll go back to the Guide pages themselves and swap back to the 8.x syntax.

I'm not certain how many I changed but it was fewer than the total number of code files that changed in the repo, not all of the guides contained page content reference the display setup portion of the code. If I had to guess I would say maybe only a 12-20 pages changed whereas 56 py code files changed.

tulip sleet
#

I would like a better annotation mechanism. It could also just be comments. it makes the code.py a bit bulky.

lone axle
#

but I also didn't mark which ones they were so I'll have to work through the list again, but it's fresh in my memory so I think I will spot them quickly this time.

tulip sleet
#

unless you took the opportunity to do other copyediting

lone axle
#

Nice I'll look for that way, I've never used it before but I assume it's on the right side controls.

#

Is there by chance a list of overall user activity on a page somewhere? then I could go thru that list of only the ones I touched in the last day or two.

tulip sleet
#

Yes, click on history. It should be relatively obvious what to do.

manic glacierBOT
#
dhalbert
[adafruit/circuitpython] New tag created: v1.21.0
#
dhalbert
[adafruit/circuitpython] New tag created: 2.2.0rc1
manic glacierBOT
idle owl
#

@tulip sleet So, apparently the Feather ESP32-S3 Reverse TFT buttons, the three on the front, D0 is false when pressed, and D1/D2 are true when pressed. Is there still a way to use keypad with this?

#

Two instances of keypad makes things sad.

tulip sleet
idle owl
#

Ah ok. I think what I was trying to do isn't happy with that, but I was printing both events at the same time, so it was printing None each time. Bad usage since that's not how it would work anyways. I think I understand. Thank you.

orchid basinBOT
manic glacierBOT
orchid basinBOT
manic glacierBOT
#
dhalbert
[adafruit/circuitpython] New review comment on pull request #9095: add RF.Guru RP2040 board

I'm going to merge in this cosmetic spacing change. We tend to group pin aliases together surrounded by whitespace to make them more obvious.

    { MP_ROM_QSTR(MP_QSTR_GP25), MP_ROM_PTR(&pin_GPIO25) },
    
    { MP_ROM_QSTR(MP_QSTR_GP26_A0), MP_ROM_PTR(&pin_GPIO26) },
    { MP_ROM_QSTR(MP_QSTR_GP26), MP_ROM_PTR(&pin_GPIO26) },
    { MP_ROM_QSTR(MP_QSTR_A0), MP_ROM_PTR(&pin_GPIO26) },
    
    { MP_ROM_QSTR(MP_QSTR_GP27_A1), MP_ROM_PTR(&pin_GPIO27) },
    { MP_ROM_QSTR(...
manic glacierBOT
tulip sleet
#

@slender iron @onyx hinge I think I will do a 9.0.1 tonight or tomorrow morning with the current set of useful 9.0.x fixes: SO_REUSEADDR, displayio warnings off, minor pin and doc fixes

slender iron
#

sounds good! thanks Dan!

misty garnet
#

@tulip sleet My pull request for minimqtt is failing because it can't import adafruit_ticks, does that need to be added to the test build environment?

tulip sleet
manic glacierBOT
#
anecdata
[adafruit/circuitpython] New comment on issue #9003: `MemoryError` with HTTPS server and `SSLSocket.accept()` on ESP32-S2, Pico W (possibly other)

On PicoW with the adafruit_httpserver library, after the MemoryError: is caught in httpserver, gc.mem_free() shows ~40KB+ free.

Tested an alternate simple-as-possible socket-level HTTPS server (w/o the library)... free memory before the accept call is ~80KB+, and ~50KB+ before the socket close. But simple requests can be received, and simple responses can be returned.

So it may be a "purely" memory issue (as opposed to something else masquerading as a memory error)... perhaps some...

fringe yarrow
#

I wonder if my issue getting an original ESP32 to work with a Parallel Display Bus LCD is down to how esp-idf seems to reference DMA and PSRAM within the i80 transmission functions..

manic glacierBOT
#
dhalbert
[adafruit/circuitpython] New tag created: 9.0.1
orchid basinBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
lone axle
manic glacierBOT
orchid basinBOT
tulip sleet
#

https://blog.adafruit.com/2024/03/26/circuitpython-9-0-1-released/
Notable changes include:

  • Implement socket SO_REUSEADDR on Pico W.
  • Turn off warnings about displayio name changes introduced in 9.0.0.
Adafruit Industries - Makers, hackers, artists, designers and engineers!
CircuitPython 9.0.1 Released!

From the GitHub release page: This is CircuitPython 9.0.1, the latest bugfix release of CircuitPython, and is a new stable release. WARNING for nRF52 boards only: If your board has an nRF52 UF2 boo…

orchid basinBOT
tulip sleet
#

I will ping our Silabs github contacts if i don't hear back sometime tomorrow.

short tendon
#

If we have our own hook, we can also do things like make rules for code that doesn't work. Like importing something from os that doesn't exist...

wraith crow
#

@tulip sleet Should routine maintenance PRs go against 9.0.x?

tulip sleet
wraith crow
#

Just looking at adding some frozen libraries to the new Sunton board, haven't tested it out yet to see if it's worth doing.

tulip sleet
midnight ember
#

Is there a reason why I shouldn't submit an example for httpserver with connection manager? Seems like Anecdata wasn't too keen on mixing them together. I have no idea what either of them does under the hood.

wraith crow
#

thx 😁

tulip sleet
midnight ember
#

ah ok yes that was the intention, only for set up.

tulip sleet
#

that is, the args passed to the constructor or whatever

midnight ember
manic glacierBOT
lone sandalBOT
midnight ember
#

Turn off warnings about displayio name changes introduced in 9.0.0. Aww I actually really liked that feature.

manic glacierBOT
#
DJDevon3
[adafruit/circuitpython] Issue opened: #9098 AnalogOut NotImplementedError on Adafruit ESP32-S3 Feather

CircuitPython version

Adafruit CircuitPython 9.0.0 on 2024-03-19; Adafruit Feather ESP32S3 4MB Flash 2MB PSRAM with ESP32S3
Board ID:adafruit_feather_esp32s3_4mbflash_2mbpsram

Code/REPL

# This used to work fine in 8.0.0 beta on an Adafruit ESP32-S2 Feather
from analogio import AnalogOut
# Adafruit arcade button with LED. Only supplies power to the LED.
Reset_LED = AnalogOut(board.A0)
Reset_LED.value = 65535

I switched boards to the ESP32-S3 feather ...
pure spade
#

Hello here, I have a question that I think belongs here more than in #help-with-circuitpython , but let me know if I am wrong.
I am adding support for a board that has an SSD1306 screen. I added adafruit_ssd1306 and adafruit_framebuf as frozen modules so that the screen is automatically usable.
My issue is that font5x8.bin is missing to draw text on the screen, and I am looking for a way to embed this file in the uf2 firmware to have it automatically available on the CircuitPython filesystem.

fringe yarrow
#

You can define which font is used as the default display font like this board definition does:

#

There's almost assuredly some other board that does what you're looking to do, or very close to it, if you can find the right search terms.

stuck elbow
fringe yarrow
#

If a driver for an as yet unsupported epaperdisplay is available for micropython, how big of an undertaking would it be to port it? Essentially a rewrite, or a mild coercion?

stuck elbow
#

depends how different the protocol is from the already supported ones

#

often it's just a question of specifying a different init sequence

fringe yarrow
#

Naw its super different. It doesn't even have a traditional controller chip. Its a 8 bit parallel wired panel that the esp32 drives directly.

stuck elbow
#

8 bit seems like too few bits for direct driving, is this just an 8-segment display?

fringe yarrow
#

in particular, it's this product: https://www.lilygo.cc/en-ca/products/t5-4-7-inch-e-paper-v2-3

LILYGO®
T5 e-Paper

LILYGO® T5 4.7 Inch E-paper V2.3 ESP32-S3 Development Driver Board Display Module Support TF Arduino Compatible Raspberry Pi Due to the problem of the characteristics of the ink screen, it cannot be partially refreshed for a long time, otherwise, it will leave residual images and lead to irreversible damage. T5 4.7 Inc

stuck elbow
fringe yarrow
#

That;'s the v1.. the v2.3 is parallel

stuck elbow
#

see second page of the schematic/

#

in any case, the fact that it uses parallel instead of spi doesn't make it direct drive

#

you just use a different bus, the driver should be the same

fringe yarrow
#

Yeah, I don't know where I read the direct drive thing, clearly 8 bits wouldn't be enough for that.. I'm fairly sure it is 8 bit parallel, though.. they, unsurprisingly, seem to have conflicting information on their data dump..

stuck elbow
#

yeah, just found it too

#

there is a datasheet for the display

#

but nothing abut the chip

#

ah, no, github lies to me, there is more

#

anyways, I think it should be relatively easy to port

manic glacierBOT
#
bablokb
[adafruit/circuitpython] New comment on issue #9085: Pico-W memory bloat

Tested this with a Trinket-M0 (SAMD21):

  • 8.0.5: 18416
  • 8.2.10: 18480
  • 9.0.1: 17552

This is in so far interesting, since we have 9.0.1/8.2.10 = 95%. And when I disable PICODVI and USB_HOST I have 94% for the Pico-W. So my assumption is that these 5% loss are from merging MicroPython. Drilling down into the symbols might give some more insights.

I also worked through the .data, .bss and .itcm sections, but there is not much room for improvement left. Disabling mdns, webworkflow an...

manic glacierBOT
manic glacierBOT
#
vladak
[adafruit/circuitpython] New comment on issue #9066: Feather ESP32 V2: Internal watchdog timer expired after switching to CP 9.0.0

The above tight loop with sensor readings has been running for 3 days contiguously (thanks to the power bank and also a bit due to the spring weather with some sunshine that charged the battery), without hitting the dreaded internal watchdog timeout. By now it has consumed both batteries so it is slowly descending towards drained battery, unless there is a lot of sunshine tomorrow which is likely, however because of these 2 properties (drained batteries and no sight in reproducing the problem...

fringe yarrow
stuck elbow
#

you should be able to get something working by taking an existing eink driver, switching it to parallel and replacing the init sequence with one from that micropython driver

fringe yarrow
#

assuming this display follows standards in place, sure..

manic glacierBOT
#
dhalbert
[adafruit/circuitpython] New comment on issue #9098: AnalogOut NotImplementedError on Adafruit ESP32-S3 Feather

An annotation should probably be added to the Essentials learn guide Analog out page to advise users that there is no DAC on ESP32-S3 and they should use PWM.

The Essentials Analog Out section was written way before non-atmel-samd boards, and could use some cleanup to be less specific. The board-specific Learn Guides for ESP32-S3 do not include an Analog Out page (or at least the ones I looked at don't).

stuck elbow
#

why wouldn't it

manic glacierBOT
#
dhalbert
[adafruit/circuitpython] New comment on issue #9085: Pico-W memory bloat

What I would do now: Pimoroni has a number of boards (mainly with e-ink displays) that have an integrated Pico-W. Unless you have advanced soldering skills, the pins are not accessible. So disabling PICODVI and USB_HOST on these boards would probably not hurt anybody but would reclaim 20K of RAM. Maybe there are other boards as well that are candidates for this change, but I don't know.

That's a fine idea, if there is no way to have those use PCIODVI and USB host in any case. Thanks!

manic glacierBOT
orchid basinBOT
tulip sleet
#

@slender iron I think it's time to do 9.1.0-beta.0, ya think? No pending PR's that look necessary. Is there something you are awaiting from TinyUSB you want to include?

manic glacierBOT
slender iron
slender iron
slender iron
fringe yarrow
#

it's hidden behind the display which is stuck down with adhesive.. but it's got 8 data pins, sth, ckh, cvk, stv and le pins..

#

as well as data, sck, adn str..

#

so lots of pins..

#

its a model ed047tc1 epaper display

manic glacierBOT
tulip sleet
#

i am going grocery shopping and will start on the release when I get back.

manic glacierBOT
slender iron
fringe yarrow
#

Yeah, I didn't think so.. It's the same panel as the M5Paper, but that has a IT8951E chip driving the display through SPI

slender iron
#

that's a different class of displays than what we already support

#

yup, we usually have a driver chip to talk to

fringe yarrow
#

i think the work entailed bringing that support to Circuitpython is beyond me at this point 😄

slender iron
#

ya, it's not simple

fringe yarrow
#

well, i guess I'll try to come up with something clever to do with this device in edf-isp / arduino..

manic glacierBOT
#
jepler
[adafruit/circuitpython] New review comment on pull request #9102: pico_dvi: support 640x240 (tested @1bpp) and 800x240 (not tested)

so you'd rather the resolution changing code looked more like this?

     if ((width == 640 && height == 480) ||
-        (width == 320 && height == 240)) {
+        (width == 320 && height == 240) ||
+        (width == 640 && height == 240)
+        ) {
         timing = &dvi_timing_640x480p_60hz;
     } else if ((width == 800 && height == 480) ||
-               (width == 400 && height == 240)) {
+               (width == 400 && height == 240) ||
+               (width == 800...
onyx hinge
#

haha nicely done discord 🙂

#

redacting my source code

manic glacierBOT
midnight ember
onyx hinge
#

@midnight ember also classic

midnight ember
#

i come back to discord and be like oh nooo it's so shouty. doesn't matter if you edit it. once it pops into dev chat it stays like that permanently. it's honestly a bit embarrassing. 🥹

stuck elbow
#

that bot should just escape markdown

midnight ember
#

i think part of the problem is that it doesn't see the backticks everywhere. sometimes it's printing them literally instead of formatting it with them.

midnight ember
fringe yarrow
orchid basinBOT
manic glacierBOT
fringe yarrow
#

Anyone ever bothered to put Circuit Python on a quadcopter flight controller? They would be an interestings platform for robotics projects, having a bunch of position and accelleration sensors and servo outputs, and such..

manic glacierBOT
idle owl
#

Hey, folks! My editor split an fstring into two lines, and CircuitPython failed with invalid syntax. I added a + to the first line and it works. Is that expected?

candid sun
#

that has been my experience

idle owl
#

Thanks so much!

slender iron
fringe yarrow
#

I've got a MatekF405 fc here. Might try to get it and all its telemetry sensors and such working in circuit Python tomorrow, just to see if I can.

manic glacierBOT
midnight ember
#

here's an example. this goes in and out of f-string mode. technically this is the most correct form of an f-string concatenation. only using f"" when/where needed.```py
TWITCH_FOLLOWERS_SOURCE = (
"https://api.twitch.tv/helix/channels"
+ "/followers?"
+ f"broadcaster_id={TWITCH_UID}"
)

also, in circuit python sometimes the above might fail with a syntax error but the below will succeed.
```py
TWITCH_FOLLOWERS_SOURCE = ("https://api.twitch.tv/helix/channels"
            + "/followers?"
            + f"broadcaster_id={TWITCH_UID}"
            )

it can be very picky about indentation with concatenated strings.

TWITCH_FOLLOWERS_SOURCE = (f"https://api.twitch.tv/helix/channels"
            + f"/followers?"
            + f"broadcaster_id={TWITCH_UID}"
            )
```while this one does work, pylint will hiss at you like an angry opossum for using f-strings in lines without {variables}
manic glacierBOT
idle owl
manic glacierBOT
manic glacierBOT
manic glacierBOT
#
dsteidley
[adafruit/circuitpython] New comment on pull request #8862: Re-enable paralleldisplaybus on ESP

When I try to use the ST7789 driver for this, I get an error:
Traceback (most recent call last):
File "code.py", line 14, in <module>
AttributeError: 'module' object has no attribute 'SPI'

The code was the example at https://github.com/adafruit/Adafruit_CircuitPython_ST7789

... and the 'module' is 'board'

I am gathering from the above conversation that LilyGo isn't using SPI to drive the ST7789. Can you help point me in the direction of how to get the CircuitPython Adafruit_gfx...

#
dhalbert
[adafruit/circuitpython] New comment on issue #9079: IDF reports resource not available on soft reboot

This is strange. I installed 9.0.1 on QT Py ESP32 Pico, and used @RetiredWizard's test program, and it's not breaking for me:

Adafruit CircuitPython 9.0.1 on 2024-03-27; Adafruit QT Py ESP32 PICO with ESP32
>>> 
soft reboot

Auto-reload is on. Simply save files over USB to run them or enter REPL to disable.
code.py output:
Hellorld! 37076
Hellorld! 8797
Traceback (most recent call last):
  File "code.py", line 7, in <module>
KeyboardInterrupt: 

Code done running.

Press a...
manic glacierBOT
manic glacierBOT
midnight ember
#

Thanks for the reminder. It's umm been a while. 😬

Capture.JPG
#

Is there a way to specify a git repo or something to keep this updated? Mine is obviously not setup to automatically update.

manic glacierBOT
#
RetiredWizard
[adafruit/circuitpython] Issue opened: #9103 Automatically run repl.py after code.py finishes without press any key prompt

I would like to use repl.py to run a "virtual repl" after code.py runs on boards that aren't connected to a host PC and have an alternate input method not supported by the core (serial matrix/i2c keyboard, touch screen, etc) but currently after code.py finishes the

Press any key to enter the REPL. Use CTRL-D to reload.

message is displayed and unless an input device supported by the core is connected Circuit Python is done.

I think the advantage of this would be that as I ...

manic glacierBOT
#
RetiredWizard
[adafruit/circuitpython] Issue opened: #9104 ESP32 Sunton board - Noise on USB Serial at startup causes occasional board freeze

CircuitPython version

Adafruit CircuitPython 9.0.0-14-gc035ade1fd on 2024-03-27; sunton_esp32_2432S028 with ESP32

Code/REPL

No Code.py

or

import sys
a = ""
while a != 'q':
    a=sys.stdin.read(1)

Behavior

When the board powers up with no code.py running, random characters frequently appear as input to the REPL, I believe if the core reads the uart device for console input at the "press any key to enter REPL" prompt and encounters a non-u...

manic glacierBOT
manic glacierBOT
manic glacierBOT
tiny peak
#

@slender iron To practice for an eventual litex/orangecrab port, I'm trying to get the litex/fomu port building on a recent compiler (gcc from the fomu-toolchain v1.6 is old).

Using Debian Bookworm's gcc-riscv64-unknown-elf to compile for rv32i works well enough, but finding a suitable standard library to link against is more challenging.

Maybe I haven't looked in the right place yet, but so far I haven't found a supported RISC-V toolchain with a pre-built newlib standard library for rv32.

It seems like people who want an rv32 standard library tend to compile their own toolchain. I'd like to avoid creating a procedure that would require instructions starting with, "First you compile gcc..." I suppose it wouldn't be as slow as compiling a kernel, but still, it wouldn't be great.

At the moment, it looks like my best bet may be to use a prebuilt compiler from Debian (or Ubuntu LTS), then use a fetch-port-submodules to download newlib source, then add a make rule to build newlib? Does that sound plausibly reasonable?

GitHub
Release v1.6 · im-tomu/fomu-toolchain
manic glacierBOT
manic glacierBOT
short tendon
midnight ember
#

That's the problem. I never remember to.

short tendon
#

I would also say it probably doesn't matter. Things don't change that much between major versions...

manic glacierBOT
tulip sleet
candid sun
#

i'm working on the sht4x trinkey guide and starting on the sht4x demos in CP. when i call i2c = board.I2C() i get ValueError: Invalid pins. when i call i2c = busio.I2C(board.SCL, board.SDA) the sensor instantiates and the loop runs. dir(board) returns ['__name__', 'I2C', 'NEOPIXEL', 'SCL', 'SDA', 'TOUCH', 'board_id']. it seems odd that board.I2C() is not working unless i'm missing something?

tulip sleet
candid sun
#

thank you!

onyx hinge
#

@tulip sleet oh yeah it would be nice if this was in 9.0.x. I'll take a look this morning and switch which branch it's going in to.

tulip sleet
#

there's a compile error and it would be good to test for regressions for the typical non 640x240 case

tulip sleet
onyx hinge
#

OK, I'll look at it still

candid sun
tulip sleet
candid sun
#

@tulip sleet that works 🙌 i was going to be putting it into review later today/early tomorrow probably

tulip sleet
#

ok, I can make a quick release, np

#

will make a PR and make you a reviewer

candid sun
#

thanks, i appreciate it

manic glacierBOT
tulip sleet
#

@onyx hinge - ok, as seen above we need a quick 9.0.2, so if you want to rebase and test for regressions in the 640x240 PR It will go in 9.0.2, which I will try to release later today.

onyx hinge
#

@tulip sleet no, probably best not to try to do this under pressure. Thanks though! I can make my project work either way, the text is just scrunched if the 200 line mode is not available.

tulip sleet
#

ok, we'll just wait on that and you can reference a 9.1.0-beta.x in any guide

manic glacierBOT
tulip sleet
#

@slender iron I was going to make a 9.1.0-beta.0 today. Only addition from yesterday is a pin fix for trinkey sht4x. Do you have anything else to go in? Same fix is also going into a 9.0.2 (only fix for 9.0.1 -> 9.0.2, due to impending guide)

lone axle
# midnight ember Is there a way to specify a git repo or something to keep this updated? Mine is...

I don't think there is a mechanism built in to update it automatically but there are a few ways that it could be set up. The easiest is probably using Task Schedule (on Windows) or Cron (on Linux) or whatever the Mac equivalent is, to schedule a script to run once per day or however often you want have have it run a command like pip install circuitpython-stubs --upgrade I think that will check for versions newer than what you've got and install them if found. So if you run that once per day you'd stay up to date with new releases.

#

Another option perhaps more complex is to clone the circuitpython core repo locally and then use a similar type of daily repeating script to pull the repo and then build the stubs. Then you could use "editable" install with pip by going into the stubs directory and doing pip install -e . that will basically dynamically install whatever is in the directory you run it in (stubs in this case) and it will automatically notice any new changes in that directory. So if the next day you build stubs again and there is new stuff it will automagically be recognized in your python environment. This method is really meant more for development, and the way that the stubs need to be built out of the core repo ends up adding extra steps to this process. I'd probably recomend method above with install --upgrade instead. But if you ever find yourself working on local changes to something installed in pip this one is very convenient during development.

manic glacierBOT
#
dhalbert
[adafruit/circuitpython] New tag created: 9.0.2
tulip sleet
#

The only change to 9.0.2 from 9.0.1 is a fix for a pin error for the Adafruit SHT4x Trinkey board. You do not need to upgrade from 9.0.1 unless you are using this board.

manic glacierBOT
#
dhalbert
[adafruit/circuitpython] Pull request review submitted: #9105 serial.c filter uart console non utf-8 on startup

We have seen similar junk on UART lines on startup, caused by other hardware issues: https://github.com/adafruit/circuitpython/issues/8890.

If this only happens on startup, then I think the fix is to clear the input buffer early on to discard any junk input.

I am not sure if your fix will work when a UTF-8 prefix is read and no further character is read immediately. Will the prefix then be discarded? If someone is trying to send a multi-byte UTF8 character then this I think may prevent ...

tidal kiln
#

any known issues with ADC on nRF52?

Adafruit CircuitPython 9.0.1 on 2024-03-27; Adafruit ItsyBitsy nRF52840 Express with nRF52840
>>> import board
>>> import analogio
>>> adc = analogio.AnalogIn(board.A1)
>>> adc.value
0
>>> adc.value
0
>>>

reporting 0 regardless of actual value (using 10k pot as voltage divider)
same behavior on all Ax pins

tulip sleet
tidal kiln
#

@tulip sleet thanks. fyi. 8.2.10 works. so seems to be a 9.x thing.

# 
Adafruit CircuitPython 8.2.10 on 2024-02-14; Adafruit ItsyBitsy nRF52840 Express with nRF52840
>>> import board
>>> import analogio
>>> adc = analogio.AnalogIn(board.A1)
>>> adc.value
35826
>>> adc.value
18785
>>>
tulip sleet
#

maybe I'll need to release 9.0.3 today 🙂

slender iron
slender iron
#

for exactly that reason

manic glacierBOT
#
RetiredWizard
[adafruit/circuitpython] New comment on pull request #9105: serial.c filter uart console non utf-8 on startup

I thought I remembered seeing the issue before but I guess I didn't look hard enough. I like your idea of clearing the input buffer.

I think this PR only addresses a problem caused waiting for the "press any key to enter repl" prompt and it looks to me like that prompt only reads a single byte and doesn't really do anything with it except maybe check for ctrl-D. I don't understand how the non-utf8 char was causing the board to freeze but I haven't been able to reproduce the freeze since ma...

wraith crow
#

@danh I'm fine closing both #9104 and #9105 in favor of #8890, what do you think?

tiny peak
tulip sleet
#

I am not sure it is the same bug: 8890 happens on ESP32-S3, but not -S2. I did not test on plain ESP32. But the probem might also be due to the USB-serial converter chip on the board not being reset properly, etc.

slender iron
tulip sleet
#

in your case, I think some kind of clearing of the input buffer would help. I often see dozens of junk characters as first input on these ESP32 boards. There is often no reset line available for the USB-serial converters

tiny peak
slender iron
#

I went even further and was experimenting with moving away from makefiles

#

though it is way more practical to integrate into our existing makefiles

wraith crow
#

Okay, I'll spend some time trying to figure out how to do a buffer clear, I may scream for help at some point 😁

orchid basinBOT
manic glacierBOT
manic glacierBOT
#
RetiredWizard
[adafruit/circuitpython] New comment on issue #9103: Automatically run repl.py after code.py finishes without press any key prompt

I'm doing this on something like the cardputer or Hacktablet not connected to a host so all input goes though a virtual keyboard on the tablet or the cardputer keyboard.

The device boots into PyDOS through an import in code.py. The idea is that if PyDOS is exited or crashes a python coded version of the REPL would be available and started via repl.py. I'll look at set_next_code_file, I've tried using that for launching programs from PyDOS that wouldn't fit in memory along with PyDOS altho...

tulip sleet
tidal kiln
#

@tulip sleet probably no surprise, but same thing happening on CPB

Adafruit CircuitPython 9.0.1 on 2024-03-27; Adafruit Circuit Playground Bluefruit with nRF52840
>>> from adafruit_circuitplayground import cp
>>> cp.temperature
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "adafruit_circuitplayground/circuit_playground_base.py", line 650, in temperature
  File "adafruit_thermistor.py", line 126, in temperature
  File "adafruit_thermistor.py", line 116, in resistance
ZeroDivisionError: division by zero
>>>
#

the denominator is just an analog read of board.TEMPERATURE

tulip sleet
#

I think the compiler may be over-optimizing something

tidal kiln
#

@tulip sleet want me to open an issue for general tracking?

tulip sleet
#

i'll open one, that's fine, since I am already on the track of the cause. I'd kind of like to fix this pretty quickly, maybe today.

#

since it's pretty bad

manic glacierBOT
#
dhalbert
[adafruit/circuitpython] New tag created: 9.1.0-beta.0
orchid basinBOT
tulip sleet
#

https://blog.adafruit.com/2024/03/28/circuitpython-9-0-2-released/
The only change to 9.0.2 from 9.0.1 is a fix for a pin error for the Adafruit SHT4x Trinkey board. You do not need to upgrade unless you are using this board.

Adafruit Industries - Makers, hackers, artists, designers and engineers!
CircuitPython 9.0.2 Released!

From the GitHub release page: This is CircuitPython 9.0.2, the latest bugfix release of CircuitPython, and is a new stable release. The only change to 9.0.2 from 9.0.1 is a fix for a pin error for …

manic glacierBOT
#
dhalbert
[adafruit/circuitpython] Issue opened: #9109 9.0.x: nRF: AnalogIn.value always returns 0

Starting with early 9.0.0 releases, AnalogIn on nordic boards returns zero all the time, no matter what input is applied to the pin. For instance, even with A0 tied to 3.3V, you see:

>>> import analogio,board
>>> a0 = analogio.AnalogIn(board.A0)
>>> a0.value
0

The underlying code has not changed in some time.

We are currently using gcc13. I recompiled a nordic build 9.0.x with gcc10, and the problem goes away (!).

orchid basinBOT
manic glacierBOT
orchid basinBOT
tulip sleet
#

https://blog.adafruit.com/2024/03/28/circuitpython-9-1-0-beta-0-released/

Notable changes since 9.0.0:

  • Audio
    • RP2040 I2SOut supports BLCK and LRCLK in either order.
  • Built-in modules
    • Enhance collections.deque functionality.
    • Add keypad_demux.DemuxKeyMatrix: use multiplexer for one side of a keypad matrix.
    • Add integration-based debouncing to keypad.
  • Internal
    • Espressif: update to ESP-IDF v5.2.1.
Adafruit Industries - Makers, hackers, artists, designers and engineers!
CircuitPython 9.1.0 Beta 0 Released!

From the GitHub release page: This is CircuitPython 9.1.0-beta.0, a beta for CircuitPython 9.1.0, and is a new unstable release. This release has known bugs that will be addressed before 9.0.0 fina…

manic glacierBOT
#
seanritner
[adafruit/circuitpython] Issue opened: #9111 Pico W UF2 missing core modules on 9.0.2

CircuitPython version

Adafruit CircuitPython 9.0.2 on 2024-03-28; Raspberry Pi Pico W with rp2040
Board ID:raspberry_pi_pico_w

Code/REPL

>>> help("modules")

Please wait a moment while I gather a list of all available modules...

/Applications/Thonny.app/Contents/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages/_distutils_hack/__init__.py:33: UserWarning: Setuptools is replacing distutils.
  warnings.warn("Setuptools is replacing...
#
todbot
[adafruit/circuitpython] New comment on pull request #8862: Re-enable paralleldisplaybus on ESP

Yep, @dsteidley I just verified what @SeanTheITGuy said is correct: just installing the CircuitPython for the LilyGo TDisplay S3 automatically creates the display and you can get it for your own purposes with display = board.DISPLAY.

If you do want to create your own, it looks like the equivalent is something like the below. But what I'd like to know:
why is rowstart & colstart swapped and rotation=270 compared to [lilygo_...

manic glacierBOT
#
RetiredWizard
[adafruit/circuitpython] New comment on pull request #9105: serial.c filter uart console non utf-8 on startup

@dhalbert I tried moving the read loop down well below the board_init call hoping it would resolve #8890 but no matter where I tried the S3 I was testing on continued to come up with the stray byte. If you have any thoughts as to where the S3 needs the buffer cleared, let me know and I'll give it a try.

With the loop where I have it now seems to resolve the issue on the Sunton board.

wraith crow
#

@tulip sleet I'm not real happy with either of the Sunton UART fixes, filtering for non-UTF characters in serial.c still passes noise to the REPL so it feels partial at best and adding a buffer clear loop to main.c just to mitigate what's probably a device reset bug in the hardware design doesn't seem like a good idea either. Unless you think there's a better solution I'm thinking I should just close the issue as a hardware design flaw and be done. I'm also thinking the S3 bug feels like it's something different, I don't think that one happen in 8.x.

tulip sleet
manic glacierBOT
#
RetiredWizard
[adafruit/circuitpython] New comment on issue #9104: ESP32 Sunton board - Noise on USB Serial at startup causes occasional board freeze

After some testing I don't think this is the same issue as #8890. My best guess at this point is that this is caused by a startup issue with the board design. If there's no code.py running to catch and discard the noise it can be annoying to start the board with the UART connected to a host, as it may take several boot attempts, but if code.py is running, it can easily discard the noise.

I'm going to close this as it seems to be hardware related as well as the associated PR #9105.

#
seanritner
[adafruit/circuitpython] New comment on issue #9111: Pico W UF2 missing core modules on 9.0.2

Thonny seems to be showing you your host computer. Can you directly connect to the Pico W serial REPL with Thonny, or over web workflow, or over USB with a different serial client?

Wait I may have just figured it out based on your comment. I was able to change the intepretor language from local Python 3 to Circuitpython and it found my device:

Adafruit CircuitPython 8.2.10 on 2024-02-14; Raspberry Pi Pico W with rp2040
>>> help ("modules")
__future__        builtins          ...
#
anecdata
[adafruit/circuitpython] Issue opened: #9112 Add separate workflow credentials for native wifi without auto-connect

Thought spurred by a review discussion in adafruit_httpserver:
https://github.com/adafruit/Adafruit_CircuitPython_HTTPServer/pull/91#discussion_r1542612641

Example native wifi code commonly uses os.getenv("CIRCUITPY_WIFI_SSID") etc. with wifi connect logic, though that implies wifi has already auto-connected. Example.

Web [workflow](https://docs.circuitpython.org/en/la...

fringe yarrow
#

I just discovered something I'm curious about. Is there a technical reason why the sdioio library has been included in the atmel-samd, broadcom and cxd56 ports, but not esp32? Or is it simply a matter of lack of demand?

slender iron
#

It is implementation I think

manic glacierBOT
#
dhalbert
[adafruit/circuitpython] Issue opened: #9113 os.getenv("A_NUMBER") can return an int

If settings.toml contains:

A_NUMBER = 123

os.getenv("A_NUMBER") will return the int 123.

This is non-standard compared with CPython, where os.getenv() always returns a string or None.

This came up because I was thinking about adding TOML bool support (true and false) to what we can parse in settings.toml.

I'm not sure this is worth fixing, but I think it's possible to fix.

crimson ferry
#

this behavior seems unusual:```
Code done running.
MemoryError:
MemoryError:
MemoryError:

tulip sleet
#

does it repeat? it's so out of memory it doesn't even have room to set up the error message to rpint

crimson ferry
#

just the three times

#

but in the past I don't recall getting multiple MemoryError:

#

and getting it after code stops

tulip sleet
#

it can be debugged if you want to open an issue

crimson ferry
manic glacierBOT
#
dhalbert
[adafruit/circuitpython] Pull request opened: #9114 Fix always zero AnalogIn on nordic
  • Fixes #9109.

This is quite mysterious. The underlying code hasn't changed in quite a while. nordic AnalogIn.value was always returning zero on 9.x.x builds, including some early betas. The problem starts appearing when gcc13 starts being used. Compiling with gcc10 or gcc12 causes the problem to go away.

While debugging, I added a printf which also caused the problem to go away. Adding tiny delay calls almost anywhere also helped. Eventually I realized it may have been due to these ...

midnight ember
manic glacierBOT
#
DJDevon3
[adafruit/circuitpython] Issue opened: #9115 Adafruit ESP32-S2 Feather ESPIDF Memory & Socket Errors

CircuitPython version

Adafruit CircuitPython 9.0.0 on 2024-03-19; Adafruit Feather ESP32S2 with ESP32S2
Board ID:adafruit_feather_esp32s2

Adafruit CircuitPython 8.2.5 on 2023-09-07; Adafruit Feather ESP32S2 with ESP32S2
Board ID:adafruit_feather_esp32s2

Code/REPL

import gc
import traceback
import os
import time
import wifi
import ssl
import socketpool
import adafruit_requests

start_mem = gc.mem_free()
print(f"Prior to Script Configurations: ...
manic glacierBOT
jolly roost
#

Hi there, we are running into a strange error and can't find any solution so far. We implement CP for the senseBox MCU with ESP32S2 (https://github.com/adafruit/circuitpython/tree/main/ports/espressif/boards/sensebox_mcu_esp32s2). The standard qwiic ports use GPIO39 and 40 for the I2C. The onboard MPU6050 is connected to the secondary i2c using the SDA=GPIO45, SCL=GPIO42. If we try to setup the secondary i2c using i2c = busio.I2C(scl=board.SCL1, sda=board.SDA1) we are running into the error: ValueError: Invalid pins . We already tried the script (https://learn.adafruit.com/circuitpython-essentials/circuitpython-i2c) identifing the i2c and possible combinations, but SDA1 (GPIO45) never shows up in the list. I already tried to update the pin definitions to G42 and G45 and build a new CP but same error. Is the GPIO45 somewhere is defined, so that it is blocked? In Arduino everything is working fine.

manic glacierBOT
solar whale
#

@tulip sleet -- a nit-pick, but in the release notes shouldn'tThis is CircuitPython 9.1.0-beta.0, a beta for CircuitPython 9.1.0, and is a new unstable release. This release has known bugs that will be addressed before 9.0.0 final. read This is CircuitPython 9.1.0-beta.0, a beta for CircuitPython 9.1.0, and is a new unstable release. This release has known bugs that will be addressed before 9.1.0 final. ?

midnight ember
#

good catch.

tulip sleet
manic glacierBOT
tulip sleet
#

I will check that this is true on ESP32-S2. Right now this is not conditionalized for any particular chip. how are you doing pullups on 45 and 42?

#

From the ESP32-S2 technical reference manual:

By default, GPIO0, GPIO45 and GPIO46 are connected to internal pull-up/pull-down during chip reset.
Consequently, if the three GPIOs are unconnected or their connected external circuits are high-impedance, the
internal weak pull-up/pull-down will determine the default input level of the strapping pins (see Table 54).

#

GPIO45 is used to select the VDD_SPI power supply voltage at reset:
• GPIO45 = 0, VDD_SPI pin is powered directly from VDD3P3_RTC_IO via resistor RSP I . Typically this
voltage is 3.3 V.
• GPIO45 = 1, VDD_SPI pin is powered from internal 1.8 V LDO.

manic glacierBOT
#
anecdata
[adafruit/circuitpython] New comment on issue #9115: Adafruit ESP32-S2 Feather ESPIDF Memory & Socket Errors

I see espidf.MemoryError: on ESP32-S2 during requests _get_socket (8.x.x), even after a fresh reset. My code automatically retries a few times, and if those fail, resets (microcontroller.reset). It may go a few cycles like that, then be successful either after the retries or after one or more resets. I have suspected something variable not quite ready in the ESP-IDF on startup, though sometimes it can happen well into a run. No idea if this is comparable to your situation.

manic glacierBOT
#
jepler
[adafruit/circuitpython] New comment on pull request #9102: pico_dvi: support 640x240 and 800x240 (tested @1bpp)

I tested the following modes:

    {"width":320, "height":240, "color_depth":8},
    {"width":640, "height":240, "color_depth":1}, # (new)
    {"width":640, "height":480, "color_depth":1},
    {"width":800, "height":240, "color_depth":1}, # (new)
    {"width":800, "height":480, "color_depth":1}):

the 400x240@8bpp mode would not initialize:

MemoryError: memory allocation failed, allocating 110412 bytes

This allocation is the framebuffer of 96000 bytes plus the "TDMS ...

#
RetiredWizard
[adafruit/circuitpython] New comment on issue #8658: "could not complete raw paste" on SeeedStudio XIAO-ESP32C3

This is not a solution by any means but if you're keen on getting something working.... I have been able to use these board by uploading files using a serial transfer program similar to the way mpremote works. It's a little unfinished as I got it working "well enough" and left it at that. The UI is a bit weird but once you get used to it, works ok.

https://github.com/RetiredWizard/Serial-File-Manager

manic glacierBOT
orchid basinBOT
jolly roost
tulip sleet
#

I can give you a hacked build that allows you try out the I2C_SDA and i2C_SCL lines, but it might complain there are no pullups

jolly roost
#

yeah I just tried this and get the errors for missing pullups

tulip sleet
tulip sleet
#

did they supply some sample code that uses IO45 to read from the IMU?

jolly roost
manic glacierBOT
tulip sleet
manic glacierBOT
tulip sleet
manic glacierBOT
tulip sleet
#

maybe I missed something in the schematic, but I don't know where the pullups are coming from for the IMU. If you measure the resistance pins 42 and 45 to the +3.3v line when the box is unconnected to power, what do you see?

orchid basinBOT
manic glacierBOT
jolly roost
manic glacierBOT
tulip sleet
#

we can make an exception for this particular board in the pin check

jolly roost
tulip sleet
#

I find it mysterious that the Arduino program works, since I can't see it setting the internal pullups anywhere, but maybe it is deeper down in the code. I'll look further

orchid basinBOT
tulip sleet
#

I'm thinking it has to do with the pin's reset state which is not changed when the I2C peripheral is enabled

orchid basinBOT
manic glacierBOT
manic glacierBOT
#
anecdata
[adafruit/circuitpython] New comment on issue #9112: Add separate workflow credentials for native wifi without auto-connect

My preference is to change the examples.

Change the examples to:

  1. assume auto-connect, and remove wifi.radio.connect logic?

  2. don't assume auto-connect, and keepwifi.radio.connect logic:

2a. continue to use the settings.toml web auto-connect credentials?
(potentially confusing or undesired per OP)

2b. use alternate credentials as discussed above?
(a standard method will make support easier)

2c. implement the CIRCUITPY_AUTO_CONNECT as discussed above?

3...

manic glacierBOT
orchid basinBOT
manic glacierBOT
#
anecdata
[adafruit/circuitpython] New comment on issue #9112: Add separate workflow credentials for native wifi without auto-connect

I think @tannewt is suggesting (2b.) above. We could even devise a simple standard connect code block with comments about auto-connect vs. manual connect, and I think we could even make the code for native wifi and esp32spi more similar (would require supporting esp32spi connect with 2 params rather than / in addition to a dict¹, and adding a connected alias). Snippets borrowed from connection manager examples. Something like:

# "If the keys CIRCUITPY_WIFI_SSID and CIRCUITPY_WIF...
manic glacierBOT
manic glacierBOT
manic glacierBOT
#
UnexpectedMaker
[adafruit/circuitpython] New comment on issue #8658: "could not complete raw paste" on SeeedStudio XIAO-ESP32C3

I've also been experiencing this on my TinyC6 and had to use a modified version of the Serial-File-Manager by @RetiredWizard to be able to communicate with my C6 boards running CP when under test on my test jigs.

It would be great if we could work out what this is as it seems to be specific to the Espressif serial UART setup in their C3/C6 chips, so will likely effect the H2, C61 and C5.

manic glacierBOT
#
DJDevon3
[adafruit/circuitpython] New comment on issue #9115: Adafruit ESP32-S2 Feather ESPIDF Memory & Socket Errors

It's possible that's why I stopped using it on the S2 around 8.0 beta. It would work for a while then crash. At the time it was suspected that there was an ESP-IDF memory leak. Very likely why I haven't used this project since. At the time the ESP32-S2 was far more stable than the S3. The S3 is now so stable that it works like a dream for most of my projects.

I can switch to an S3 but that's not really the point. Yes Anecdata the behavior you're seeing is similar. It's random and t...

orchid basinBOT
#
jepler
[adafruit/circuitpython-org] New comment on pull request #1378: Generate absolute URLs automatically

generation looks good

  <!-- (uses an absolute URL for better compatibility with RSS readers -->
  <img srcset="https://circuitpython.org/assets/images/boards/small/raspberry_pi_pico.jpg 300w,
               https://circuitpython.org/assets/images/boards/large/raspberry_pi_pico.jpg 700w" sizes="(max-width: 1024px) 700px,
              300px" src="/assets/images/boards/large/raspberry_pi_pico.jpg" alt="Image of Board" loading="lazy">

            ```
orchid basinBOT
small comet
#

Silly question, but just in case: correct to assume that rp2pio isn't present in Blinka or any other library? (Context: was trying to fix CI issues in adafruit-circuitpython-pio-uart, but any library that imports rp2pio seems like it would have a problem.)

manic glacierBOT
#
RetiredWizard
[adafruit/circuitpython] Issue opened: #9119 Web Workflow Serial screen on Full Editor page crashing on non-focus input

CircuitPython version

Adafruit CircuitPython 9.0.0-beta.0 on 2024-01-27; Adafruit Feather ESP32S3 4MB Flash 2MB PSRAM with ESP32S3
9.0.2
9.1.0 beta 0

Code/REPL

a = ''
while a != 'q':
    a = input('_$ ')

Behavior

soft reboot

Auto-reload is on. Simply save files over USB to run them or enter REPL to disable.
code.py output:
_$ Traceback (most recent call last):
  File "code.py", line 3, in 
EOFError: 

Code done running.

P...
manic glacierBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
lone axle
#

Does anyone know what might lead to the readme badges being stacked vertically instead of horizontally?

I've seen this randomly on a few library repo's now. I briefly looked into it one day but could not spot any syntax differences in the readme file. I'm sure it's something I overlooked, but I'm not exactly sure what things could cause it beyond something like extra \ns inserted or similar.

image.png
#

Hmm, maybe they are all that way actually. I'm pretty sure there were still some horizontal ones when I looked into it maybe a week or two ago. but now I'm having trouble finding any examples of horizontal ones.

#

okay, something very strange IMO. When I view the readme section at the "home" URL of that repo, they are horizontal as depicted in that screenshot. But if I click on the readme file specifically they change to vertical ⁉️

midnight ember
#

what is a readme.rst? I've only seen readme.md... should I be using .rst instead?

lone axle
midnight ember
#

ah thank you. i probably could have googled that 🤦

#

detective foamyguy tracking down /n's

#

i'm seeing the same results you are.

#

some of the headings are double spaced, some are not. it looks random from manual typing input not something that's been run through adabot or formatted in a specific way.

#

wiznet looks uniformly formatted, very nicely. the PIO_UART is not uniform.

small comet
lone axle
small comet
#

Ack - thanks @lone axle !

manic glacierBOT
#
RetiredWizard
[adafruit/circuitpython] New comment on issue #8658: "could not complete raw paste" on SeeedStudio XIAO-ESP32C3

Since the C3 was on my mind, I pulled one out for a project I was working on yesterday. I did use ampy to get the initial settings.toml file loaded, but once it was connected to WiFi, I used the Web Workflow for everything. The File Browser to transfer files, the Internet based full code editor and associated internet based serial interface (not the CP hosted serital console) and everything went very smoothly. I was even able to use the Web Workflow feature of Circup to check and update the l...

#
RetiredWizard
[adafruit/circuitpython] Issue opened: #9121 Create host based app to access web workflow file management features

it might be nice if there was an application similar to Circup that could be used to manage the file system of Circuit Python devices over a Web Workflow channel.

In addition to streamlining simple operations, this might allow for automated configuration procedures to be used in setting up multiple devices or the use of operating scripts to perform more complicated file management operations.

I'll have to look back at @FoamyGuy's Web Workflow/Circup streams to see if there's a similar a...

manic glacierBOT
#
bill88t
[adafruit/circuitpython] Issue opened: #9122 ESP-IDF 5.2.1 not identifying itself as 5.2, but 5.1 instead

The following was run with ~/.espressif wiped.

[bill88t@lunchbox | ~/git]> git clone https://github.com/adafruit/circuitpython cptest && cd cptest/ports/espressif && make fetch-port-submodules && cd esp-idf
 [...]
[bill88t@lunchbox | ~/git/cptest/ports/espressif/esp-idf]> git status
HEAD detached at 722fba96a1
nothing to commit, working tree clean
[bill88t@lunchbox | ~/git/cptest/ports/espressif/esp-idf]> ./install.sh
Detecting the Python interpreter
Checking "python3" ...
P...
brazen hatch
#

I still haven't tested CP with idf5.2, did anyone check wifi authentication messages? The changelog reported a lot of changes there, and I think scanning was fixed in this release too.

manic glacierBOT
#
niccit
[adafruit/circuitpython] Issue opened: #9123 Web Workflow Impacts MQTT message flow

CircuitPython version

Adafruit CircuitPython 9.0.2

Code/REPL

import wifi
import adafruit_connection_manager
import adafruit_minimqtt as my_mqtt

cert_file = mqtt_data["cert_file"]
with open(cert_file, 'r') as file:
    cert_data = file.read()

radio = wifi.radio
pool = adafruit_connection_manager.get_radio_socketpool(radio)
ssl_context = adafruit_connection_manager.get_radio_ssl_context(radio)
ssl_context.load_verify_locations(cadata=cert_data)
...
#
bill88t
[adafruit/circuitpython] New comment on issue #9122: ESP-IDF 5.2.1 not identifying itself as 5.2, but 5.1 instead

The name for the python_env is ESP-IDF's choice, not ours.

Cloning espressif's esp-idf release 5.2.1, and running ./install.sh, shows it correctly identifying itself as 5.2.
The issue is with the forked idf, that is used for CircuitPython.

I don't understand what you mean here.

Up until now it has been possible to have a seperate esp-idf clone, elsewhere on the system and have it exported to path.
CircuitPython would then build just fine, so long as version matches.
That I r...

manic glacierBOT
ornate breach
#

which version of the arm toolchain does circuitpython use?

#

10?

stuck elbow
manic glacierBOT
manic glacierBOT
orchid basinBOT
manic glacierBOT
#
raquo
[adafruit/circuitpython] New comment on issue #7902: ItsyBitsy M4 Express deep sleep with external wakeup

This is marked as a bug, but... is it really a bug?

According to https://github.com/adafruit/circuitpython/pull/5425 the only pins that can be used for PinAlarm are "IN0:PB00; IN1:PB02; IN2:PA02; IN3:PC00; IN4:PC01". Of those, it seems that only the PA02 pin is brought out on ItsyBitsy m4 – according to the pinout chart, it is the A0 pin.

PinAlarm does in fact work on my ItsyBitsy m4's A0 pin with this code:

import alar...
manic glacierBOT
#
studiostephe
[adafruit/circuitpython] Issue opened: #9124 Web Workflow: no circuitpython.local (Pico W, 9.0.1)

CircuitPython version

Adafruit CircuitPython 9.0.1 on 2024-03-27; Raspberry Pi Pico W with rp2040

Code/REPL

#code.py (provided by anecdata in the discord)
import time
import os
import wifi
import ipaddress
import socketpool
import mdns

MDNSFINDTIMEOUT = 5

ssid = os.getenv('WIFI_SSID')
password = os.getenv('WIFI_PASSWORD')

pool = socketpool.SocketPool(wifi.radio)

def connect():
    # to connect in case CP8 web workflow hasn't already, or ...
manic glacierBOT
#
somenice
[adafruit/circuitpython] New comment on issue #9115: Adafruit ESP32-S2 Feather ESPIDF Memory & Socket Errors

I can report similar behaviour switching from 8 to 9 on Feather ESP32 S2
I can successfully requests.get() an https address but fail with IO send_data() using IO_HTTP

Adafruit CircuitPython 9.0.2 on 2024-03-28; Adafruit Feather ESP32S2 with ESP32S2

`Traceback (most recent call last):
File "code.py", line 313, in <module>
File "adafruit_io/adafruit_io.py", line 624, in send_data
File "adafruit_io/adafruit_io.py", line 564, in _post
File "adafruit_requests.py", line 595, in...

manic glacierBOT
#
UnexpectedMaker
[adafruit/circuitpython] New comment on issue #8658: "could not complete raw paste" on SeeedStudio XIAO-ESP32C3

I've been doing some thorough testing of Ampy with my TinyC6 on my test jig setup - and I can confirm the issue is still present, but it's it succeeds more often than it fails.

90% of the time it works and I can PUT files on the TinyC6.
70% of the time I can RUN a file on the device

The rest of the times a RUN hangs and a PUT fails, or the device locks up serial comms after a task completes.

This is way better than what I have been experiencing with mpremote and shell wh...

manic glacierBOT
manic glacierBOT
#
UnexpectedMaker
[adafruit/circuitpython] New comment on issue #8658: "could not complete raw paste" on SeeedStudio XIAO-ESP32C3

As there is some variety of very valuable in terms of makers chips, perhaps it is worth considering creating a new thread topic? Or maybe change the name to this one?

Well, the C3 and C6 both use the same UART peripheral, so whatever works or doesn't work for one, will be the same for the other, and all new chips from Espressif that use the same peripheral, like the C5, C61 and I believe the H2 + H4 will also have the same issues, so I think it's relevant token this here and not part mul...

#
DJDevon3
[adafruit/circuitpython] New comment on issue #9115: Adafruit ESP32-S2 Feather ESPIDF Memory & Socket Errors

Here's a revised test script that requires no logins for (8.2.5). Anyone with an S2 should be able to run this script.

Adafruit CircuitPython 8.2.5 on 2023-09-07; Adafruit Feather ESP32S2 with ESP32S2
Board ID:adafruit_feather_esp32s2

import gc
import os
import time
import wifi
import ssl
import socketpool
import adafruit_requests

print("-" * 40)
start_mem = gc.mem_free()
print(f"Prior to Script Config: {start_mem}")

# Initialize WiFi Pool (There can be...
#
DJDevon3
[adafruit/circuitpython] New comment on issue #9115: Adafruit ESP32-S2 Feather ESPIDF Memory & Socket Errors 
Adafruit CircuitPython 9.0.0 on 2024-03-19; Adafruit Feather ESP32S2 with ESP32S2
Board ID:adafruit_feather_esp32s2

code.py output:
----------------------------------------
Prior to Script Config: 2039824
----------------------------------------
After Script Config: 2038848
Memory Used: 976
----------------------------------------
Prior to WiFi: 2038848
✅ Wifi!
----------------------------------------
After Wifi: 2038848
Memory Used: 0
------------------------...
manic glacierBOT
manic glacierBOT
#
anecdata
[adafruit/circuitpython] New comment on issue #9124: Web Workflow: no circuitpython.local (Pico W, 9.0.1)

I've seen similar behavior on raspberrypi. Often, the browser tries to load but does not return with the page. The mdns server shows up in mdns.Server .find (done from a separate espressif device), but loading in a browser is problematic. In rare cases it will load in the browser (though I haven't been able to do that today), but in many of those the web workflow page will come up with all fields empty. Even accessing web workflow using the IP address sometimes shows similar behaviors...

manic glacierBOT
#
bill88t
[adafruit/circuitpython] Issue opened: #9125 IDF5.2.1 connecting to secure network without a password

CircuitPython version

Adafruit CircuitPython 9.1.0-beta.0 on 2024-03-28; DFRobot FireBeetle 2 ESP32-S3 with ESP32S3

Code/REPL

import wifi
wifi.radio.connect("secure_network")

Behavior

Traceback (most recent call last):
File "", line 1, in
ConnectionError: Unknown failure 210

Description

Now trying to connect to a secure network with no password produces code 210.

Additional information

I opened this issue to discuss as to what s...

manic glacierBOT
small comet
manic glacierBOT
manic glacierBOT
orchid basinBOT
fringe yarrow
#

So is the main branch supposed to be coming with idf5.2.1 now, or do you have to somehow install the toolchain manually? i've pulled and fetched-modules and esp-idf/install.sh export.sh and it's still 5.1

tulip sleet
orchid basinBOT
manic glacierBOT
#
dhalbert
[adafruit/circuitpython] Pull request opened: #9126 handle new ESOP-IDF 5.2.1 auth failure codes
  • Fixes #9125

ESP-IDF added some new failure codes for wifi connections. Handle a couple of these by collapsing them to existing "authentication failure" code so they will print with a useful message.

Test: old and new:

Adafruit CircuitPython 9.1.0-beta.0 on 2024-03-28; Adafruit Feather ESP32-S3 Reverse TFT with ESP32S3
>>> import wifi
>>> wifi.radio.connect("MYLOCALNETWORK")
Traceback (most recent call last):
  File "", line 1, in 
ConnectionError: Unknown failure 210
>>...
fringe yarrow
#

@danh Not sure if you remember it, but there was a mysterious "IO8" in the esp32-c3 w/ round lcd that I submitted earlier in the week.. I just found it.. It's a tiny solder pad next to the UART connector,, I gues so you could hook a switch or a neopixel to it..

tulip sleet
manic glacierBOT
#
anecdata
[adafruit/circuitpython] Issue opened: #9127 raspberrypi Out of MDNS service slots

CircuitPython version

Adafruit CircuitPython 9.1.0-beta.0 on 2024-03-28; Adafruit QT Py ESP32-S3 4MB Flash 2MB PSRAM with ESP32S3

vs.

Adafruit CircuitPython 9.1.0-beta.0 on 2024-03-28; Raspberry Pi Pico W with rp2040

Code/REPL

import time
import os
import traceback
import microcontroller
import wifi
import socketpool
import mdns

PORT1 = 8086
PORT2 = 8088

time.sleep(3)  # wait for serial
print(f'{os.uname()=}')
print(f'{microcontroller.c...
manic glacierBOT
#
anecdata
[adafruit/circuitpython] New comment on issue #8149: Out of MDNS service slots

Footnote: it does not appear possible to have the same MDNS service-protocol on different ports:

service_type and protocol can only occur on one port. Any call after the first will update the entry’s port.

A non-MDNS HTTP Server can be run in code.py with web workflow enabled, but not an MDNS HTTP Server. User code could also (temporarily to permanently) commandeer the web workflow _http._tcp service-protocol to its own port for an in-code.py MDNS HTTP Server, but at that point, web ...

manic glacierBOT
orchid basinBOT
#
jepler
[adafruit/circuitpython-org] Pull request opened: #1383 Continue trying to fix RSS feed absolute URLs

Originally, both the RSS feed and the board list pages used the same board_image.html include; I added | absolute_url to this file for the RSS feed, which functions better when all links are absolute.

Using | absolute_url is preferable to just prepending https://circuitpython.org because it can give a correct URL when previewing locally with bundle run jekyll serve.

This crossed with a change by @makermelissa to switch the feeds to use a separate include, presumably (though I di...

#
jepler
[adafruit/circuitpython-org] New comment on pull request #1383: Continue trying to fix RSS feed absolute URLs

Yes, unless I'm failing to understand something or my testing is insufficient, the "feed_board_image.html" from this issue include would work equally well for the board list pages from the point of view of a user.

In 'view source' you would see a redundant specification of the full URL of the images (img src="https://circuitpython.org/...") so in a sense it's "less tidy", but it works the same.

If you want me to re-do this PR so that it re-combines the two include files I'm happy to do ...

#
makermelissa
[adafruit/circuitpython-org] New comment on pull request #1383: Continue trying to fix RSS feed absolute URLs

If you want me to re-do this PR so that it re-combines the two include files I'm happy to do that. My assumption had been that since you split them out subsequent to my two PRs adding |absolute_url the first time that either you were encountering a problem with the version I had written, or you didn't want the redundant URLs in view-source.

Please do. I was just in a hurry as I was at the end of my day and it was the simplest way to make the change that I knew would work for cert...

orchid basinBOT
onyx hinge
#

surprise <@&356864093652516868> -- I'm hosting the meeting today in about 80 minutes. Hope to see you in the voice channel & your notes in the notes doc.

orchid basinBOT
manic glacierBOT
tiny peak
#

re: voice chat roles... um.. I'm cool with just listening, but I could also say hi

onyx hinge
#

listening in for your first time is a solid idea. we're happy to add you to the role anytime. it also gets you a few notifications a week, just about the meeting times.

#

cool that you've joined us in any case!

thorny jay
#

Lurking without notes...

manic glacierBOT
#
tannewt
[adafruit/circuitpython] New comment on issue #9115: Adafruit ESP32-S2 Feather ESPIDF Memory & Socket Errors

One second I have 2.1MB of memfree then immediately afterward on the next line 1 single call to requests.get throws a memory error. It doesn't seem like a lack of memory or at least not according to the behavior reported by memfree. It's not like I have 10KB free... I have around 2MB free when it crashes. The API call I'm making when it crashes most times should only take about 200 bytes...

memfree is telling you how much is available for MP heap. It isn't accounting for the IDF memory r...

lone axle
#

https://mastodon.social/@todbot/112197138537048342

mastodon.social
Tod Kurt (todbot)
Tod Kurt (todbot) (@todbot@mastodon.social)

Attached: 1 video

There’s a neat new core library in CircuitPython: bitmapfilter. Originally designed for the PyCamera project, it’s got many useful operations for generative art stuff like I do in Processing/p5js. Here’s a start: random circles bluring away. note, it does require a fast processor like the ESP32-S3 as on this great Lilygo boa...

▶ Play video
crimson ferry
#

sorry, text only

manic glacierBOT
idle owl
lone axle
slender iron
lone axle
#

I do like autoconnect personally because it's nice to not have that portion of boilerplate in user code for some of my projects and examples that I test.

thorny jay
#

I don't fully get everything, but what about password less wifi, if something is trigger by the presence of a password?

lone axle
#

something like CIRCUITPY_WIFI_AUTOCONNECT being a way to turn it on /off would be nice I think as well but don't have a strong opinion on it.

thorny jay
#

Sorry to have confused things, but now I get it. 😉

manic glacierBOT
#
niccit
[adafruit/circuitpython] New comment on issue #9123: Web Workflow Impacts MQTT message flow

Yes, perfectly. My project had been running since Saturday without any
weirdness. And being super clear, WWF is still enabled. I just haven't
accessed since Saturday.

On Mon, Apr 1, 2024, 11:36 Scott Shawcroft @.***> wrote:

Does MQTT work ok when web workflow isn't enabled?

—
Reply to this email directly, view it on GitHub
https://github.com/adafruit/circuitpython/issues/9123#issuecomment-2030323059,
or unsubscribe
<https://github.com/notifications/unsubscr...

crimson ferry
#

for examples: if we assume web workflow enabled, do simple examples even need connect logic?

dense glacier
#

wifi.radio.enabled = False
Would this prevent autoconnect?

crimson ferry
#

...and, when connect logic is desired (but no auto-connect), do we want some pseudo-standard settings.toml credentials to use and recommend?

#

@dense glacier auto-connect would briefly connect on startup, then stop once that line gets hit in code

idle owl
#

Thanks for hosting, Jeff! I need to head out. Have a great week, everyone!

dense glacier
#

Thanks all

turbid radish
#

thanks!

gilded cradle
#

Thanks

lone axle
#

Thanks for hosting Jeff, hope everyone has a great week!

manic glacierBOT
#
anecdata
[adafruit/circuitpython] New comment on issue #9124: Web Workflow: no circuitpython.local (Pico W, 9.0.1)

CIRCUITPY_WEB_API_PASSWORD= "jk"

I think this is the web workflow case. My understanding is that mdns.Server(wifi.radio) in code accesses what's essentially the web workflow mdns.Server Singleton?

(Running the code above I don't think will show its own device, but running on a separate device should catch any MDNS _circuitpython._tcp.)

@studiostephe can chime in on their original observations from Discord. From my own observations, it can be difficult to connect to `raspberrypi...

lone axle
#

Would it be considered a bug (i.e. should issue be created) the EPaperDisplay.time_to_refresh (https://docs.circuitpython.org/en/latest/shared-bindings/epaperdisplay/index.html#epaperdisplay.EPaperDisplay.time_to_refresh) is innacurate by a partial second? The docs state "Time, in fractional seconds, until the ePaper display can be refreshed." but I'm finding pretty consistently if I call refresh() as soon as that time is 0.0 I get the "Refresh Too Soon" error.

I ended up catching the error and working around the issue by adding some extra time "manually" with:

time.sleep(display.time_to_refresh + 0.6)
display.refresh()
manic glacierBOT
#
niccit
[adafruit/circuitpython] New comment on issue #9123: Web Workflow Impacts MQTT message flow

If you'd like a video or to see it over a zoom/hangout call I'm happy to do
that.

20 years in QA and I know how important a picture can be with weird issues
like this. :)

Code:

~nicci

On Mon, Apr 1, 2024 at 11:45 AM Nicci Tynen @.***> wrote:

Yes, perfectly. My project had ...

spare jacinth
#

i guess the code could do something like time_since_last_draw >= (display.time_to_refresh - some_delta) to make it a bit friendlier in that regard (at the cost of not complying 100% with datasheet/recommendations)
but i would have expected the CircuitPy internals to automatically flush the display (if it has changed) after time_to_refresh has elapsed, and not that users need to eagerly flush themselves 🤔

#

could also store in milliseconds and use int instead? 🤔

slender iron
manic glacierBOT
#
anecdata
[adafruit/circuitpython] New comment on issue #9124: Web Workflow: no circuitpython.local (Pico W, 9.0.1)

For example, I just set up a Pico W with no code.py, but web workflow and auto-connect set up in settings.toml. I've turned off web workflow on all other devices.
Adafruit CircuitPython 9.1.0-beta.0 on 2024-03-28; Raspberry Pi Pico W with rp2040

A separate MDNS scanner shows the Pico W:

  • ipaddress.ip_address(pool.getaddrinfo(circuitpython.local, 80)[0][4][0]) gives the proper IP address.
  • .find(service_type="_circuitpython", protocol="_tcp") gives the .hostname as the `cpy...
#
dhalbert
[adafruit/circuitpython] New comment on issue #9112: Add separate workflow credentials for native wifi without auto-connect

@justmobilize The meeting video is not up yet. I was multitasking during the meeting and don't remember if you attended, but we discussed this in In the Weeds today.

  • Using CIRCUITPY_WIFI_SSID and CIRCUITPY_WIFI_PASSWORD causes auto-connect.
  • Using CIRCUITPY_WEB_API_PASSWORD enables web workflow if auto-connect has happened.
  • The CIRCUITPY_ prefix on the settings.toml values is meant to signify that CircuitPython does something itself with these values.
  • There is currentl...
#
FoamyGuy
[adafruit/circuitpython] Issue opened: #9129 EPaperDisplay.time_to_refresh is too low to ensure safe refresh

CircuitPython version

Adafruit CircuitPython 9.0.0 on 2024-03-19; Adafruit MagTag with ESP32S2

Code/REPL

import time
import keypad
import board

display = board.DISPLAY
buttons = keypad.Keys((board.BUTTON_A, board.BUTTON_B, board.BUTTON_C, board.BUTTON_D), value_when_pressed=False)

while True:
    event = buttons.events.get()
    if event:
        if event.released:
            print(f"btn {event.key_number}")

            try:
              ...
#
justmobilize
[adafruit/circuitpython] New comment on issue #9112: Add separate workflow credentials for native wifi without auto-connect

@dhalbert thanks. Sadly I missed it today...

So we would probably want something like this in the examples:

if not radio.connected:
    # We got here because CIRCUITPY_WIFI_SSID/CIRCUITPY_WIFI_PASSWORD are either incorrect or not set
    # Try grabbing WIFI_SSID/WIFI_PASSWORD and connecting
    ssid = os.getenv("WIFI_SSID")
    password = os.getenv("WIFI_PASSWORD")
    while not radio.connected:
        try:
            radio.connect(ssid, password)
        except Connect...
orchid basinBOT
onyx hinge
#

Here is the notes document for next Monday’s CircuitPython Weekly Meeting. It is at the normal time of 11am Pacific / 2pm US Eastern here on Discord. Add your hug reports and status updates to the document before the meeting. If you are unable to attend but would still like to contribute, feel free to add your notes and we’ll read them off during the meeting. Hope to see you there! <@&356864093652516868>
https://docs.google.com/document/d/1_cpbdNujG6Gb6YB18n2v94Jqyz09h4yPsyiRRcy9tgk/edit?usp=sharing

Google Docs
Copy of TEMPLATE ONLY -- CircuitPython Weekly Meeting for [DATE (TE...
onyx hinge
manic glacierBOT
#
dhalbert
[adafruit/circuitpython] New comment on issue #9112: Add separate workflow credentials for native wifi without auto-connect

Video is up. Discussion starts here: https://youtu.be/fbg2qtGjxC8?si=hlf-NvkHwluSLIRg&t=1987

So we would probably want something like this in the examples:

That is the most thorough boilerplate. The retry while loop could be omitted, I think, and then also the try-except. I notice @anecdata also includes that in their example.

I think one question is whether to bother to include this for all examples or just depend on auto_connect.

Something like this code could go in `Conn...

orchid basinBOT
manic glacierBOT
proven garnet
lone axle
#

not discussed yet, but it does appear to be quite widespread now. We'll have to decide if we want to try to do something to get it back to horizontal or if the vertical stack isn't bothersome.

proven garnet
#

Strangely, not EVERY other repo does either (like externally), so I wonder why it's not consistent

#

I tried looking into it last week as well on a personal repo and I couldn't manage to fix it, though that was where I confirmed that the RTD documentation was unaffected

#

We had a similar issue a while back where it wouldn't line them up horizontally without a :target: attribute I think, but that doesn't seem to be the case this time around.

spare jacinth
# manic glacier

@lone axle took a quick look before going to bed and AFAICT, the C code already works in ms but it gets converted when "handing" the value to Python with

shared-bindings/epaperdisplay/EPaperDisplay.c
return mp_obj_new_float(common_hal_epaperdisplay_epaperdisplay_get_time_to_refresh(self) / 1000.0);

#

and that's probably causing some rounding errors, as i was already guessing before

lone axle
#

feel free to leave a comment on the issue with your findings.

spare jacinth
#

aight 👌

lone axle
#

And thanks for looking into it!

manic glacierBOT
#
elpekenin
[adafruit/circuitpython] New comment on issue #9129: EPaperDisplay.time_to_refresh is too low to ensure safe refresh

AFAICT the C code works in milliseconds yet it gets converted to seconds when "passing" it to Py, which probably causes a rounding error which is the real problem making the code fail

I think a viable solution would be to do a breaking change so that the property returns in milliseconds and call sleep_ms instead of sleep

On a side note, maybe there should be some (opt-in) ...

spare jacinth
#

🛏️ 💤 👋

manic glacierBOT
#
anecdata
[adafruit/circuitpython] New comment on issue #9112: Add separate workflow credentials for native wifi without auto-connect

In any robust application, there would be a lot of exception-handling, etc. But for simple examples, there are two cases:

Auto-Connect enabled:
No wifi.radio.connect() logic is needed, but some documentation about settings.toml and auto-connect would set the stage.

Auto-Connect not enabled:
We could make it super-simple and simply call wifi.radio.connect(), no ifs or whiles even needed. .connect() is virtually a NOP if already connected, there is very little overhe...

lone axle
#

Are there any devices that have an analog capable pin that has a resistor of a known value connected to it? I've been kicking around an idea in the back if my mind to try to catch things like that AnalogIn issue linked above.

It would be cool if we could make a utility that runs on a PC but connects to a connected CircuitPython device either via Filesystem and Serial, or HTTP and Websocket. It could load specific script(s) and run them testing for the output to match the expected values.

It probably can't easily comprehensively test everything but it could do quick sanity checks on the most basics like digitalio, analogio and maybe busio by communicating with a known connected I2C device or something.

If such a device / pin exists with a resistor connected internally then one script could init the pin and read the value to compare with the expected value based on that built-in resistor in order to do a baseline validation that analogio is functional.

short tendon
lone axle
#

Nice, that seems like another nice usage of that type of mechanism.

tulip sleet
lone axle
#

I ended up down a similar road, I connected 12 to 13 and then read 13 while setting 12

tulip sleet
#

to really answer your question, on some chips you can enable a pull-up or pull-down even when the pin is in analog input mode. But that would just give you the ends of the range

lone axle
tulip sleet
#

you could also put some kind of bidirectional analog switch on multiple pins

tulip sleet
#

doing it on an RPi at the time was not stable enough. Lots of USB troubles and then also the RPI SD cards went bad too quick. A NUC or other small computer makes more sense. Now I have a lot of these because I'm already on my second generation of buying them.

#

(Dell Optiplex micro form factor, e.g.)

lone axle
#

When I get to writing I'll start something fresh. I'm imagining starting with sort of "one shot" tests run from the command line without any other automation built in. Eventually this could get used by some other utility to automate it but I have in mind just the manual tool to start.

manic glacierBOT
#
dhalbert
[adafruit/circuitpython] New comment on issue #9088: Memory Allocation Fail with Neo Trinkey

I reproduced this with 9.0.2 and 9.1.0-beta.0.

I tried this recent change to MicroPython, which allows some compilation garbage to be collected more reliably: https://github.com/micropython/micropython/pull/14136.

If I comment out the last four or so routines in TOS3.py (code.py), then I don't get the storage problem. So we are just over the edge.

If I use mpy-cross on TOS3.py, morse.py, and prt.py, replace the .py files with those .mpy files, and replace code.py with:
`...

manic glacierBOT
orchid basinBOT
crimson ferry
#

weird thing... several of my ESP32-S2 devices, which I haven't touched and have been running since 7.2.0 (early-2022), suddenly all get espidf.MemoryError (_get_socket) when trying to connect to server ...something changed in TLS protocol or certs or something like that? (they do reload or reset as necessary when errors occur, but can't get past this now) ... 7.3.3 devices with almost identical code.py are unaffected

orchid basinBOT
manic glacierBOT
crimson ferry
manic glacierBOT
midnight ember
# lone axle When I get to writing I'll start something fresh. I'm imagining starting with so...

Have you looked into the tests that adafruit runs on boards before putting them into inventory? There's a tests repo that Lady Ada mentioned in last weeks Desk of Lady Ada. Adafruit's test rigs mostly use fiducials and pogo pins but it seems like they might be pivoting to a new type of solderless header (coming to the store soon) https://www.youtube.com/watch?v=f1RlHjUq2iE By using the same headers that a user would, it will allow users to run the Adafruit test process on their board in the event of an incident like you're experiencing. I certainly like the idea of being able to factory test my boards if I suspect an issue. We have factory reset .bins but no factory test suite, at least not something that is beginner friendly with a learn guide for all boards... yet.

tulip sleet
midnight ember
#

ohh, welp nvm then.

#

the idea is a good one though.

tulip sleet
#

we have something like that for RPI, I haven't looked at the new ones in detail yethttps://www.adafruit.com/product/3662

Hammer Header Male - Solderless Raspberry Pi Connector

If your soldering isn't quite up to scratch, or you just don't own a soldering iron yet, then these nifty hammer headers from Pimoroni might be just what you ...

#

these are harder to use because there are so many to insert, so you need a lot of force. Using a "hammer" jig helps

midnight ember
#

I think a board test suite could help support staff in the forums too. That would be nice if they could give someone a piece of code that would test everything on the board.

#

I think most physical issues end up in the forums or email before going to RMA?

tulip sleet
#

I started working on some straightforward test programs for vetting CircuitPython board defs (not for factory use) but it didn't get that far.

midnight ember
#

You would have to emulate for analog/digital input and probably a bunch of other things.

#

I'd like to chime in on the discussion for web workflow/wifi auto connect. One of the things I like about the current boilerplate code in code.py for connecting is if wifi or dns goes down. It needs logic to reconnect and retry forever for an API script that is to be online 24/7. As long as that can still be handled I'm ok with whatever the devs decide.

#

Most of my projects are online 24/7 and require a good amount of try/except error handling to keep them online.

#

In an ideal world especially for beginners, code.py should not require a single try/except.

manic glacierBOT
#
DJDevon3
[adafruit/circuitpython] New comment on issue #9115: Adafruit ESP32-S2 Feather ESPIDF Memory & Socket Errors

I didn't think it would be since this one seems to have continually affected the S2 since its debut. memfree reporting everything is good to crashing without warning is a bit jarring.

Having a TLS buffer error would be welcome so at least users will know it's TLS buffer related and not due to lack of PSRAM. After looking into it more yesterday and testing it in different ways all HTTP sites can be sequentially loaded without issue. It's specifically HTTPS responses that causes this on...

manic glacierBOT
orchid basinBOT
manic glacierBOT
willow totem
#

I was keen to do a test farm too, but thought probably worth trying wokwi before setting up some monstrous local farm. You can point pins at pins and have simple components attached which would cover a lot of the test surfaces (but never be capable of all on wokwi)

manic glacierBOT
#
George-TL
[adafruit/circuitpython] Issue opened: #9130 usb_hid.set_interface_name() is not working

CircuitPython version

Adafruit CircuitPython 9.0.2 on 2024-03-28; Raspberry Pi Pico with rp2040

Code/REPL

# in boot.py file
usb_hid.set_interface_name("HP")

Behavior

USP iInterface field does not change.

Description

According to pull request https://github.com/adafruit/circuitpython/commit/ffa45dedb1ea306c5419e872b021f6da02f9b7bb usb_hid.set_interface_name() should work under 9.0.2.

Additional information

@MariuszCwikla what CP versi...

midnight ember
#

Pylint hates me.

manic glacierBOT
#
SeanTheITGuy
[adafruit/circuitpython] Pull request opened: #9131 Makergo esp32c3 supermini

This PR adds support for the Maker Go ESP32C3 Supermini, a very cost effective and prolific, if basic, esp32c3 dev board. It closely resembles several other boards, but UART configuration was an issue getting extant Circuit Python variants to work on it. This PR is tested working on real hardware.

Draft until pending PR to add device to the creationid database is completed.

orchid basinBOT
midnight ember
austere acorn
midnight ember
#

the struggle is real

#

At least on my S3 from typing does not import therefore the check for TYPE_CHECKING will probably always be false and from circuitpython_typing won't get imported either. ```py
try:
from typing import TYPE_CHECKING, List, Tuple, Optional, Iterable
from PIL import Image

if TYPE_CHECKING:
    from circuitpython_typing import ReadableBuffer

except ImportError as e:
pass

# 
# This takes precedence over image() in __init__ and is tuned for the
    # Matrix class. Some shortcuts can be taken because matrix layout is
    # very straightforward, and a few large write operations are used
    # rather than pixel-by-pixel writes, yielding significant speed gains
    # for animation. Buffering the full matrix for a quick write is not a
    # memory concern here, as by definition this method is used with PIL
    # images; we're not running on a RAM-constrained microcontroller.
```  Except I am running it on a RAM-constrained microcontroller...  🤷
#

Not the image function specifically but the library yes.

#

I have no idea how to resolve this to make both Blinka and microcontroller circuit python happy with typing.

lone axle
#

I initial aims with device testing aren't to have a comprehensive test that checks every pin or tries to validate every possible thing on the device. Though I think that kind of thing would certainly be valuable too.

My primary goal at least to start is to have the tests cover enough to be confident that the build of CircuitPython is okay. So things like analogio being broken for one port would get caught, but things like a specific bad pin on a specific device, or even a bad bin as a result of a bug in the PCB design would not necessarily get caught unless you were "lucky" enough to try the exact right pin(s) during the test.

#

The actual thing that I have in mind is more the "test runner" than the tests themselves honestly though. I am imaginging a tool that lets you run a command like:

circuitpython_hardware_test digitalio_test.py digitalio_test.exp

Where the first argument is the script to run and the second is a plaintext file containing the expected output of the script.

That mechanism could then apply quite generally to tests of different purposes.

#

Doing in wokwi would be cool as well. I'm not opposed to that path at all, I just don't think I have the knowledge to push it forward as much.

midnight ember
#

That's a neat idea. Use the power of Wokwi's emulation to mock up some tests.

fringe yarrow
#

so, completely unhinged thing I've been working on, I've hit a wall.. I've got this device [https://www.mateksys.com/?portfolio=f405-std] which has a STM32F405RGT6 MCU and a bunch of sensors and flash and such attached to it.. I have a fairly detailed pin arrangement, as defined in the INAV (flight control software) target for this device [https://github.com/iNavFlight/inav/blob/master/src/main/target/MATEKF405/target.h]. However, I'm missing some critical piece of information about the STM port, because trying to get a custom CircuitPython build to boot on it gives no joy. I've tried using the console from the vcp (which I may not have defined properly) as well as when hooked to the UART1 pins with an external usb serial device, and defining UART1 as the console UART in mpconfigboard. I get no output.. I'm not sure if it's not booting, or if it's just not putting out serial output, but I suspect it's not booting.. I used the feather_stm32f405 as a starting point, modified it for this devices' pins and flashed..

midnight ember
#

I feel like I need to be using Union but unsure how or why.

manic glacierBOT
midnight ember
#

How can I import ReadableBuffer and WriteableBuffer from circuitpython_typing? Whatever I'm doing isn't working.

# 
__init__.py:122:25: E0601: Using variable 'WritableBuffer' before assignment (used-before-assignment)
__init__.py:133:56: E0601: Using variable 'ReadableBuffer' before assignment (used-before-assignment)
manic glacierBOT
#
MariuszCwikla
[adafruit/circuitpython] New comment on issue #9130: usb_hid.set_interface_name() is not working

@George-TL are you testing it on windows or on linux?
[linux] can you try running lsusb -vd 239a:80f4 | grep iInterface ?
[windows] I have noticed that windows caches name of the HID device. So if you have plugged in your Pico some time ago, then it will store name in registry under HKEY_CURRENT_USER\System\CurrentControlSet\Control\MediaProperties\PrivateProperties\Joystick\OEM\VID_239A&PID_80F4 and basically "ignores" what you set up in boot.py (VID_239A&PID_80F4 are vid/pid of Ra...

hidden laurel
#

I'm using a ESP32-S3-DEV board that has native USB and USB-to-Serial convertor on UART0. I want to recompile CP so the REPL console listens on UART0 instead of the native USB created comport. I openend the board's code but I can not find where it defaults to the native USB for the REPL console.

prime flower
lone axle
spare jacinth
#

not a big deal as i can just ignore them, but is there a good reason/easy fix for submodules saying they have modified content when i have git checkout repo/main + make fetch-tags + make fetch-all-submodules?

#

did i miss some updating step? 🙃

lone axle
#

The easy fix I tend to go with when my repo starts to have weird git / submodule behavior is to just make a fresh clone and start over in there. I'm not sure it qualifies as a fix all the way, but it is easy.

orchid basinBOT
spare jacinth
orchid basinBOT
onyx hinge
# midnight ember How can I import ReadableBuffer and WriteableBuffer from circuitpython_typing? W...

I think that wrapping this import with if TYPE_CHECKING is incorrect. The following snippet fails on standard python, and is also tagged in a similar way by pylint: ```py
import typing
if typing.TYPE_CHECKING:
import typing.Union

def f(x: Union[list, str]) -> int:
return len(x)


```$ python badtyping.py 
Traceback (most recent call last):
  File "/home/jepler/badtyping.py", line 5, in <module>
    def f(x: Union[list, str]) -> int:
             ^^^^^
NameError: name 'Union' is not defined

The import of typing.Union should simply not be guarded by "if typing.TYPE_CHECKING".

spare jacinth
#

100 ms should not be a user-noticeable delay, but im hoping it is good enough to avoid rounding errors

prime flower
#

@lone axle I have a CI question if you have time.. I'm making a quick mock driver for the CircuitPython_BME280 library. It shares the same file structure as CircuitPython_BME280 where there's no expected file that follows the adafruit_drivername/drivername.py pattern. However, the CI is erroring on a FileNotFoundError

midnight ember
tulip sleet
spare jacinth
lone axle