circuitpython-dev

[thorny jay]

Waiting for a die shot jigsaw puzzle...

[lone axle]

https://pip.pypa.io/en/stable/news/

[slender iron]

https://www.youtube.com/watch?v=UH0A2iujtY8

https://github.com/sponsors/kattni

[idle owl]

Thanks! 😊

[atomic summit]

Don't forget my TinyPICO Nano, just to double down on cute words that get over used 🙂
https://www.tinypico.com/tinypico-nano

But the PICO in my case is not for "SMALL" but for the PICO-D4 on the board.

[idle owl]

@atomic summit Ergo, the need for Nano. 😄

[jaunty juniper]

I'm eagerly awaiting the itsy bitsy tiny pico nano

[atomic summit]

Exactly! A smaller than pico PICO! 😜

pico^2 😉

[solar whale]

Makes me think of "jumbo shrimp"

[atomic summit]

I missed how this conv started though - I assume the rPI Pico?

[marble hornet]

"CDC" ?

thank you.

[idle owl]

@atomic summit Originally last week sometime. The vaguely confusing branding of the board as Pico, but the chip is RP2040. And then other PIco-named boards. And then Adafruit is carrying another Pico-named board so it came up again today.

[ionic elk]

Somebody should port Circuitpython to ShenzhenIO

[thorny jay]

So I was comparing apple and apple... that's a change!

Let's say I tested the PIO code found in the assembler library for doing Neopixel.

[idle owl]

I reviewed, indeed. 🙂

[lone axle]

Thank you!

[idle owl]

You're welcome! I need to test it, but wanted to get that review in first.

[blissful pollen]

I may not be able to stay on to hear the bus device stuff, but I'll for sure listen to the recording later or someone can ping me here.

[slender iron]

I'm planning on adding a .spi property to SPIDevice

it's needed for sdcard

[fossil gorge]

Thanks for that info @idle owl. Will look at structuring the library that way, and will PR at will!

[slender iron]

https://github.com/sponsors/kattni

[mental nexus]

https://cutoutfoldup.com/981-torus-from-villarceau-circles.php

If you want to practice on your patience or “build character”...

[idle owl]

@fossil gorge Thanks! You can do a WIP PR, simply mention that it's a work in progress, and open the discussion. Bear in mind that breaking changes are allowed, but we need to be aware of what documentation such as Learn guides that need to be updated along with an eventual PR merge.

[slender iron]

wild!

https://github.com/adafruit/circuitpython/pull/3454

[solar whale]

no good deed goes unpunished!

[thorny jay]

Thanks.

[mental nexus]

Thanks all!

[modern wing]

Thanks for an awesome meeting -- I had a work call in the middle of it, so I'll have to catch the VoD 🙂

Pretty short today, done about an hour into it.

[slender iron]

Here is the notes document for Monday’s CircuitPython Weekly meeting. It is at the normal time of 11am Pacific / 2pm Eastern here on Discord. Everyone is encouraged to attend! Please add your hug reports and status updates even if you’ll be attending the meeting - it’s super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and we’ll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/1RTjIOUogdNgKSO5TdfzvsEoJBb1ohP5yREz6ZulzORA/edit?usp=sharing

[manic glacier]

eddieespinal

[adafruit/circuitpython] New comment on issue #4108: ESP32\-S2: Getting this error while running the SD card demo 'SPIDevice' object has no attribute 'spi'

@tannewt you were right, that was the problem. After trying again using the branch 6.1.x everything works fine now.

Thanks for solving my problem.

[manic glacier]

dhalbert

[adafruit/circuitpython] Pull request review submitted: #4018 add uart hardware flow control to nrf port

This looks good to me now. Your own testing vets it.

eternaltinkering

[adafruit/circuitpython] Issue opened: #4111 pulseio module missing

Adafruit CircuitPython 6.2.0-beta.1 on 2021-01-27; Raspberry Pi Pico with rp2040

import simpleio
Traceback (most recent call last):
File "", line 1, in
File "simpleio.py", line 34, in
ImportError: no module named 'pulseio'

dhalbert

[adafruit/circuitpython] New comment on issue #4111: pulseio module missing

There is no pulseio yet for CircuitPython on the RP2040. simpleio imports pulseio, but usually you just need simpleio for the map_range() function it defines. You can copy that function and add it to your program for now.

Is this an Adafruit-supplied example? We can change the example.

eternaltinkering

[adafruit/circuitpython] New comment on issue #4111: pulseio module missing

Oddly enough the simpleio module is used as an example for import errors in the getting started tutorial, mostly wanted to make sure I understood why I was getting the error after adding simpleio.mpy to the lib folder on the device.
https://learn.adafruit.com/welcome-to-circuitpython/circuitpython-libraries#example-importerror-due-to-missing-library-2978247-20

[idle owl]

@tulip sleet Oi, maybe we should choose a library that does not import other modules/libraries. Is that even a thing?

thinks

Even neopixel requires pixelbuf or pypixelbuf.

The example would need to work the same on all boards.

Helper library maybe.

[gloomy shuttle]

@idle owl is a specific widget used to embed source code in the learn system or is it something custom for Adafruit?

[idle owl]

@gloomy shuttle The entire Learn system is basically proprietary software. I'm not sure what they used. It uses the "embed" element in the Learn system software, and renders the GitHub code, as well as the links to view/download/copy it.

[gloomy shuttle]

Thanks! I kind of figured that, but it one of the best embed code displays I have seen.

[idle owl]

@gloomy shuttle If you'd like, I can ping Learn dev to find out if it was internal code or whether they implemented something existing.

[gloomy shuttle]

I would be interested to find out if it is not proprietary.

[idle owl]

@gloomy shuttle I'll ping them, then. I'll let you know what I find out.

[prime flower]

@ionic elk Thanks for unblocking me with the socket timeout core bug, just PR'd the library which was blocked 🥳

[idle owl]

@prime flower That was a BEEFY PR. Nicely done.

[prime flower]

Started in November 2020 🙂

[idle owl]

Way to stick with it!

[prime flower]

I want to use it 🙂

[idle owl]

I know that feeling 😄

[prime flower]

though, it's not fully done - I need to add SSL sockets for CPython and a test suite so I don't need to test 4 different implementations.

[idle owl]

Have to start somewhere though, oi. And yeah testing would save you a lot of time!

[prime flower]

very fun way of soak testing 🙂

[idle owl]

Nice!

[lone axle]

@gloomy shuttle if you are intersted in an open source widget for embedding code into web pages I've used this one: https://codemirror.net/ in a work project and found it to be pretty nice to work with and has a nice looking UI with many desirable features.

[ionic elk]

@prime flower wait I did? I thought I didn't have that fix up until SSLSocket goes in

[tulip sleet]

Maybe neopixel is a good example because just adding it might still generate another error (when no _pypixelbuf), so the reader could see the step-by-step process.

[idle owl]

@tulip sleet Yeah maybe. Except also not all boards have a NeoPixel so maybe not so great. hmph.

[tulip sleet]

i was just thinking of making a new library, like adafruit_simplemath, which would include map_range(), and maybe constrain(). Not sure about the naming, but it kind of parallels adafruit_simpleio.

[gloomy shuttle]

Thanks, I will check that out!

[idle owl]

@tulip sleet I'm annoyed we have simpleio to begin with.

I've written the map range stuff into my code to avoid using it.

[tulip sleet]

or maybe just adafruit_math, but that's pretty open-ended. ... yeah, I would like to substitute something simpler

[idle owl]

Repeatedly.

[prime flower]

@ionic elk https://github.com/adafruit/circuitpython/issues/3836 seems to have fixed it

er https://github.com/adafruit/circuitpython/pull/3854

I just subscribed to the SSLSocket PR you opened

[ionic elk]

3854 probably had the fix but further issues solved in 3836 might have borked it

3854 had a bunch of issues like making the timing default wrong and sockets timing out even when successful if the timeout was 0

[prime flower]

MiniMQTT doesn't rely on the timing default, it's set up by the application itself. Not sure if timeout is set to 0 anywhere though... dont think so

[ionic elk]

anyway, good to hear it's working for you now

[drowsy geyser]

Just a data point. The BNO085 on a Raspberry Pi 4 running the example code from the guide prints one line and throws an index out of range error (latest library). It's been running on the Pico for three days without a hiccup - latest CircuitPython and library bundle. Nice! At the same time, I have a 055 running on a CM4 and it's been chugging along for ten days.... Good stability.

[idle owl]

@drowsy geyser Huh! Did you file an issue on the 085 lib? Good that the others are working well!

[prime flower]

@ionic elk Yeah, I think I've (somehow) avoided the issues but I'll leave that PR up for a day

[ionic elk]

There shouldn't be any now, 3836 is in

[drowsy geyser]

I haven't yet, @idle owl . I will shortly. I wanted to test on the Pico to make sure it wasn't a hardware issue.

[ionic elk]

SSLSocket is coming but it doesn't have anything related to your issue (I got it mixed up)

[idle owl]

Good plan. Thanks ahead of time for filing an issue!

[prime flower]

I'll likely need to implement the SSLSocket interface when it's merged in. :\

@ionic elk https://docs.python.org/3/library/ssl.html#ssl-nonblocking << btw, if you're going for CPython behavior, maybe see if nonblocking sockets can be handled like in CPython.

nonblocking ssl sockets are weird in Cpython (imo)

[drowsy geyser]

Now I need to be better about following up on them!

[idle owl]

Hah, valid. But life happens, we understand.

[ionic elk]

@prime flower SSLSocket has an identical interface. No changes should be required.

[prime flower]

ok!

[ionic elk]

as for nonblocking, I'm already trying to do that, I don't know how successfully haha. You're welcome to bring up divergent cases with me and I'll try and fix them

Some non-blocking stuff is easy, like recv and accept - they just fail if they're going to block, and you gotta try them over and over until they work. Those should already be working as intended. But other stuff I haven't figured out yet, like connect, which returns an exception, and you're supposed to use select to deal with it, which I haven't touched at all. I've just been defaulting to blocking on cases like that

I guess documenting those inconsistencies would probably be good

at least in the comments, or something

[tidal kiln]

@drowsy geyser that pi+bno08x combo is pretty tricky. it's another clock stretcher. maybe some help here:
https://forums.adafruit.com/viewtopic.php?f=19&t=174424

[manic glacier]

jfabernathy

[adafruit/circuitpython] New comment on issue #4108: ESP32\-S2: Getting this error while running the SD card demo 'SPIDevice' object has no attribute 'spi'

I get this same error on both ESP32-S2 and Raspberry PI pico, but works on Seeeduio Xiao with same version of CP

[manic glacier]

gamblor21

[adafruit/circuitpython] Pull request opened: #4112 Fixed for boards without longint

Fix to avoid using any long int for builds that have it disabled. The end keyword is now also omitted if the default value of INT_MAX is given to avoid converting it to a smaller value.

Tested on a Feather M4 Express with long int disabled and a BME280.

gamblor21

[adafruit/circuitpython] New comment on issue #3856: adafruit\_bus\_device\-in\-core crashes with non\-busio objects

Fixed in #3936

gamblor21

[adafruit/circuitpython] Issue closed: #3856 adafruit\_bus\_device\-in\-core crashes with non\-busio objects

[manic glacier]

jerryneedell

[adafruit/circuitpython] Pull request review submitted: #4112 Fixed for boards without longint

Tested on hallowing_m0_express by running lis3dh_simpletest -- worked fine.

[idle owl]

@gloomy shuttle I heard back from my devs. The actual display of the GitHub embedded code is simply using Google Prettify, wrapped around a container with the header that shows it's a GitHub embed and the links to it, etc. To get the code and keep it in sync, they use the GitHub API and save it within our system for performance reasons. Then they have webhook integration that notifies us if the files change via commits, and if they do, update the files in the Learn database. (If also notifies if a file is moved and allows us to alert the authors to the situation, as handling files that have moved is super tricky and prone to failure in Learn.)

Not sure if that helps, but that's what I got. 🙂

[gloomy shuttle]

Thank you, that helps a lot! I was just looking at the Google Prettify after I originally asked.

[idle owl]

@gloomy shuttle Excellent!

[fossil gorge]

So trying to commit some changes in the Adafruit_CircuitPython_ProgressBar, and I'm running afoul of the pre-commit hook. Technically speaking, existing code is, but I'm working to correct the issues.

What's failing right now is the "reuse" hook, which checks for copyright and license information for every file. Every file. This includes .github/workflows/* files, .pylintrc, favicon.ico, requirements.txt, docs/* etc.

To make sure I get it right, what would the license and copyright information be for most of these files, if not all? Some have info in them, such as "(c) 2020 Brent Rubell for Adafruit Industries", and some mention either the Unlicense or MIT. However, all files that don't, well... don't!

[idle owl]

@fossil gorge If no one else gets to this, I'll be able to help you tomorrow. But it's getting late here and I'm done for the night.

[fossil gorge]

Sounds good, thanks! Have a good night!

[slender iron]

@fossil gorge what files are missing it? the cookiecutter should have licenses for everything it adds

[fossil gorge]

It's complaining about these files...

The following files have no copyright and licensing information:
* .github/workflows/build.yml
* .github/workflows/release.yml
* .pylintrc
* .readthedocs.yml
* CODE_OF_CONDUCT.md
* README.rst
* docs/_static/favicon.ico
* docs/api.rst
* docs/conf.py
* docs/examples.rst
* docs/index.rst
* examples/progressbar_magtag_simpletest.py
* examples/progressbar_simpletest.py
* requirements.txt
* setup.py

The following files have no licensing information:
* adafruit_progressbar.py

[slender iron]

hrm, do you have a docs/api.rst.license file? There is one in cookiecutter

https://github.com/adafruit/cookiecutter-adafruit-circuitpython/tree/master/{{ cookiecutter.library_name }}/docs

[fossil gorge]

There is not one, no.

[slender iron]

when did you run cookiecutter?

[fossil gorge]

I didn't. This is an existing library: https://github.com/adafruit/Adafruit_CircuitPython_ProgressBar

[slender iron]

ah, that'd be why

[fossil gorge]

Initial commit was... 13 months ago

[slender iron]

kattni and dherrada would know how to update it. or you can try copying the newer versions of files from cookiecutter

[fossil gorge]

Just added that one file, and reuse is happy. Literally!
Congratulations! Your project is compliant with version 3.0 of the REUSE Specification :-)

Smiley and all

[slender iron]

interesting 🙂

maybe it's only checking files you've changed

[fossil gorge]

No, it was checking all of them

At least running reuse lint directly. I didn't change anything under ".github", and I even tried it on a clean "main" branch.

But through pre-commit, it only did the changed ones, since pre-commit stashes anything not already added

[slender iron]

kk, well that unblocks you then 🙂

[fossil gorge]

Definitely. Thank you kindly!

[slender iron]

np

[fossil gorge]

Welp, it was happy. Then I created a branch to work from and commit into, and it's back to that error. 😕

[lone axle]

does the new branch have that file?

[fossil gorge]

yep

Comparing the files to the cookiecutter, none of those listed files have the header though

[lone axle]

Ah it seems like this one got left out of the sweep that added lots of this stuff

[fossil gorge]

Well crud, I think I saw why it was happy originally... I was on the branch where I was trying to see what was needed for reuse to pass, and not a clean branch 🤦

[lone axle]

There are quite a few changes needed I think. Here is a similar one of the PRs https://github.com/adafruit/Adafruit_CircuitPython_Display_Shapes/pull/30/files

I'd probably recommend skipping it for now. There might have been a script used that will make the process simpler

[fossil gorge]

That's what I was thinking. I'll commit locally without pre-commit running, and circle back around to the license stuff later

[lone axle]

yep. we can get progress_bar sorted. Then you'll be able to merge it in and not have to deal with it much I think.

[fossil gorge]

Except for any new files I add, because space-time, bits and bytes and such

[manic glacier]

tannewt

[adafruit/circuitpython] Pull request review submitted: #4096 Translations update from Weblate

tannewt

[adafruit/circuitpython] Pull request closed: #4096 Translations update from Weblate

[lone axle]

When you are around @idle owl and maybe @trim elm It seems like https://github.com/adafruit/Adafruit_CircuitPython_ProgressBar got left out of the run pre-commit / add license sweep. Was there a script in adabot or somewhere that was used to make those branches from the PRs?

[idle owl]

No.

That was all done manually I think, or Dylan ran something locally.

I'll have Dylan take a look when he's around next.

[manic glacier]

tannewt

[adafruit/circuitpython] Pull request review submitted: #4112 Fixed for boards without longint

Looks good to me! Thank you!

tannewt

[adafruit/circuitpython] New comment on pull request #4112: Fixed for boards without longint

Single CI failure was a network error. Merging anyway.

tannewt

[adafruit/circuitpython] Pull request closed: #4112 Fixed for boards without longint

tannewt

[adafruit/circuitpython] New comment on issue #4109: native BusDevice breaks I2C if LONGINT\_IMP = NONE

Fixed by #4112

tannewt

[adafruit/circuitpython] Issue closed: #4109 native BusDevice breaks I2C if LONGINT\_IMP = NONE

tannewt

[adafruit/circuitpython] New comment on issue #4107: Using GP15 on RPi Pico can prevent code\.py from running automatically

I suspect that code.py was working but the GP15 button didn't work. The errata fix doesn't restore the mux setting last I checked. @kilograham, is this on your radar?

tannewt

[adafruit/circuitpython] New comment on issue #4101: SocketPool\.Socket\.\_\_hash\_\_\(\) is not immutable

@hierophect Can you open the second LWIP socket only when you do the server APIs?

tannewt

[adafruit/circuitpython] New comment on issue #4094: Add optional pin pull\.UP/DOWN to countio\.Counter for open\-collector sensors

Does rotaryio provide pull-ups for encoder pins? I've been able to read encoders without pull-up resistors. Hmmm.

I don't think so. I think our guide has you hook one up.

tannewt

[adafruit/circuitpython] New comment on issue #4094: Add optional pin pull\.UP/DOWN to countio\.Counter for open\-collector sensors

Before trying countio, an internal pull-up was sufficient to work with both optical and hall-effect interrupter sensors. The pull-up value is not critical for most low to medium-speed applications like motor RPM. Don't know if the internal pull-up value provides enough current for the speed of a remote control IR sensor.

I'm ok if you want to add a kwarg for this.

tannewt

[adafruit/circuitpython] New comment on issue #4108: ESP32\-S2: Getting this error while running the SD card demo 'SPIDevice' object has no attribute 'spi'

I get this same error on both ESP32-S2 and Raspberry PI pico, but works on Seeeduio Xiao with same version of CP

I think the Xiao works because it uses the python version of bus device. The other two have it built in.

tannewt

[adafruit/circuitpython] Pull request review submitted: #4018 add uart hardware flow control to nrf port

Looks good to me too! Thanks!

tannewt

[adafruit/circuitpython] Pull request closed: #4018 add uart hardware flow control to nrf port

CedarGroveStudios

[adafruit/circuitpython] New comment on issue #4094: Add optional pin pull\.UP/DOWN to countio\.Counter for open\-collector sensors

I'm completely unsure about where to start when it comes to modifying one of the core elements. Time to look at how that's done. Thanks.

ladyada

[adafruit/circuitpython] New comment on issue #4094: Add optional pin pull\.UP/DOWN to countio\.Counter for open\-collector sensors

@CedarGroveStudios you can do it! we can help support ya, try making a draft PR when you're ready for some eyes :)

weblate

[adafruit/circuitpython] Pull request opened: #4113 Translations update from Weblate

Translations update from Weblate for CircuitPython/main.

Current translation status:

CedarGroveStudios

[adafruit/circuitpython] New comment on issue #4094: Add optional pin pull\.UP/DOWN to countio\.Counter for open\-collector sensors

Thanks! Another fun trip outside my comfort zone!

@tannewt: Interestingly, the rotaryio guides I've looked through that involve CircuitPython have no pull-up resistors for the two quadrature encoder pins. https://learn.adafruit.com/rotary-encoder/overview

tannewt

[adafruit/circuitpython] Pull request review submitted: #4110 RP2040: Support for RTC

Looks good to me! Thank you so much!

tannewt

[adafruit/circuitpython] New review comment on pull request #4110: RP2040: Support for RTC

Oooh, thank you!

tannewt

[adafruit/circuitpython] Pull request closed: #4110 RP2040: Support for RTC

[manic glacier]

tannewt

[adafruit/circuitpython] Pull request opened: #4114 Add \.spi accessor to SPIDevice

Fixes #4108

dhalbert

[adafruit/circuitpython] Pull request review submitted: #4114 Add \.spi accessor to SPIDevice

LGTM!

[trim elm]

@lone axle thanks for letting me know. I’ll take a look tomorrow afternoon

[lavish saffron]

Thank you. You've fixed it for me and my last push has passed all the checks. Much appreciated.

[manic glacier]

hathach

[adafruit/circuitpython] New comment on issue #3986: Copying multiple larger files can hang the copy if autoreload is on

There seems to be race condition with the background code, I put segger rtt for debugging. Normally after usb irq, background is added and then run_all. When the issue occurs, the irq is trigger, background is added, along with tick is also added, but there is no run_all() afterwards. background_callback_add_core: 48: repeatedly is tick timer trying to add its background while already existed in the queue. Doing more testing

usb_irq_handler: 103:
background_callback_add_core: 52:...

[manic glacier]

m4tk

[adafruit/circuitpython] Issue opened: #4115 ESP32\-S2: hangs during startup if backlight pin is used in display init code

Hi there,

I'm currently trying to write the display init code for the newly merged ESP32-S2 LilyGo T8 board with ST7789 display.

In general it's working so far. But I'm struggeling somehow with the backlight pin. If I've understood the whole thing correctly it should be enough to pass the adress of the mcu_pin_object as parameter to the common_hal_displayio_display_construct() function.

But if I do that the board is hanging at startup. If I pass NULL as backlight pin and set the ...

[manic glacier]

hathach

[adafruit/circuitpython] New comment on issue #3986: Copying multiple larger files can hang the copy if autoreload is on

update: tracing down the background callback, it does solve the race condition pretty well, however, somehow the run_code.py decide to run background_callback_reset() while background list is still queued up with usb event. https://github.com/adafruit/circuitpython/blob/main/main.c#L305

        usb_irq_handler: 103:
background_callback_add_core: 52:
callback_head = 0
callback_tail = 0
callback_head = 200029BC
callback_tail = 200029BC
background_callback_add_core: 65:

run...

waiweng83

[adafruit/circuitpython] New comment on issue #4107: Using GP15 on RPi Pico can prevent code\.py from running automatically

I tried it out and confirmed that program using GP15 does not run when saved as code.py or main.py.
It also has some bad impact on the REPL.

However, saved program run from REPL seems to just work fine.

waiweng83

[adafruit/circuitpython] New comment on issue #4107: Using GP15 on RPi Pico can prevent code\.py from running automatically

I suspect that code.py was working but the GP15 button didn't work. The errata fix doesn't restore the mux setting last I checked. @kilograham, is this on your radar?

@tannewt I tried building my own fork and playing around with the errata fix and seems like the call to this function is causing the problem.

// Force LS_J
const uint dp = 15;
//const uint dm = 16;

gpio_set_function(dp, 8);

Even I tried saving the original value prior to calling the function and restore i...

[manic glacier]

dhalbert

[adafruit/circuitpython] New comment on pull request #4078: Clarified I2C init error msg

@BiffoBear Your latest commits rolled back the ulab submodule a few commits by accident. I think this may have been due to not pulling from your own repo after I made changes to the PR, so they got out of sync. Also the circuitpython.pot may have lost a bunch of new messages.

Let's start over again to clean this up. Could you get the latest adafruit/main into your main, start a new fix branch, and then include only the error message changes in the various I2C.c files, and my shrink...

[manic glacier]

kilograham

[adafruit/circuitpython] New comment on issue #4107: Using GP15 on RPi Pico can prevent code\.py from running automatically

I suspect that code.py was working but the GP15 button didn't work. The errata fix doesn't restore the mux setting last I checked. @kilograham, is this on your radar?

i had heard some mention from someone, not sure the specifics... can you open an SDK issue

[manic glacier]

hriday111

[adafruit/circuitpython] Issue opened: #4116 Raspberry Pi Pico: NotImplementedError: UART not yet supported

this happens when I try to use:
busio.UART(board.GP0, board.GP1, baudrate=9600)

[idle owl]

@idle wharf Hey! Ping me when you're around, please.

[crimson ferry]

@ionic elk does this work for you with #4049 https://github.com/adafruit/Adafruit_CircuitPython_Requests/issues/63#issuecomment-771712906?

[analog bridge]

@slender iron how is flash partitioning done on rp2040?
I need NVM_SIZE and NVM_START_ADDR.

[slender iron]

@analog bridge it's all part of the LD script. you could take the last 4k of the 1MB I set aside

right now it's 1MB for the core and the rest for FS

[manic glacier]

ladyada

[adafruit/circuitpython] New comment on issue #4116: Raspberry Pi Pico: NotImplementedError: UART not yet supported

yep not implemented yet
please subscribe to
https://github.com/adafruit/circuitpython/issues/4036
to get updates when we've added it :)

ladyada

[adafruit/circuitpython] Issue closed: #4116 Raspberry Pi Pico: NotImplementedError: UART not yet supported

[analog bridge]

got it 👍

[manic glacier]

tannewt

[adafruit/circuitpython] Issue opened: #4117 Update pico\-sdk to 1\.0\.1

Fixes a number of things and exposes a way to get a flash's unique id: https://github.com/raspberrypi/pico-sdk/releases/tag/1.0.1

[slender iron]

MicroPython 1.14 has been tagged: https://github.com/micropython/micropython/commit/78b23c3a1f064dc19bbee68930ef3aba110c781c

[manic glacier]

ajs256

[adafruit/circuitpython] New comment on issue #4117: Update pico\-sdk to 1\.0\.1

I would like to work on this, could someone point me to what I would need to change? I see /ports/raspberrypi/sdk_config/pico/, is there anything else?

tannewt

[adafruit/circuitpython] New comment on issue #4117: Update pico\-sdk to 1\.0\.1

Epic! you'll need to change the commit referenced by ports/raspberrypi/sdk and then rebuild. Things may have changed so you'll likely need to update Makefile too. I think the linker script also changed so compare the CP copy to the sdk's default one.

[manic glacier]

ajs256

[adafruit/circuitpython] New comment on issue #4117: Update pico\-sdk to 1\.0\.1

Sorry, I'm having git troubles. I can't work on this anymore, please disregard my previous comment.

[idle wharf]

@idle owl ping

[idle owl]

@idle wharf pong!

Ok, for this PR:

https://github.com/adafruit/cookiecutter-adafruit-circuitpython/pull/84

Are you comfortable taking over it?

[idle wharf]

I have a few questions about it, and I assume its not time-driven, but yeah I'd be happy to help out with that.

That's a yes. 😉

[idle owl]

@idle wharf I am around to answer any questions. It's not time-driven, as should be obvious by the fact that it hasn't really been touched since it started.

That would be greatly appreciated, it's a feature I'd really like to see added.

We'll have to update the documentation if needed, but we'll cover that closer to merging.

[idle wharf]

My biggest question is about the Travis stuff... I assume now you would want to remove that and just do Github Actions

[idle owl]

Yes, 100%.

That might trip up folks creating community libs, but no more than Travis would have, really.

I think it's not worth keeping Travis in there.

[idle wharf]

OK... I'll have a look at during the evening this week or on Friday.
github process... should I start a new PR or can I jump in on the current pr... not had to do that before. What's the CP-Community way ?

[idle owl]

@idle wharf You should be able to grab that PR and work on it. I can help you with that process. If that ends up being too complicated for you, it's perfectly fine to start a new PR as well.

It's a little tweaky to push to an existing PR using Git, but it's doable. I have to refer to notes on how to do it every time, but it works.

[idle wharf]

You right... I wasn't thinking that I have commit privileges so yeah that makes sense

OK, consider it done.

[idle owl]

Great! Thank you so much.

[manic glacier]

tannewt

[adafruit/circuitpython] Issue closed: #4108 ESP32\-S2: Getting this error while running the SD card demo 'SPIDevice' object has no attribute 'spi'

tannewt

[adafruit/circuitpython] Pull request closed: #4114 Add \.spi accessor to SPIDevice

tannewt

[adafruit/circuitpython] New comment on issue #4094: Add optional pin pull\.UP/DOWN to countio\.Counter for open\-collector sensors

Yup, I'm wrong. IncrementalEncoder on the SAMD turns them on automatically: https://github.com/adafruit/circuitpython/blob/main/ports/atmel-samd/common-hal/rotaryio/IncrementalEncoder.c#L59-L63

tannewt

[adafruit/circuitpython] New comment on issue #3986: Copying multiple larger files can hang the copy if autoreload is on

@hathach That makes sense to me! I think we need probably need to remove VM initiated task only (not USB.) An alternative would be run all tasks while interrupts are off and then we reset before turning them back on.

tannewt

[adafruit/circuitpython] New comment on issue #4107: Using GP15 on RPi Pico can prevent code\.py from running automatically

Filed https://github.com/raspberrypi/pico-sdk/issues/76 . Thanks!

[idle owl]

@slender iron Does https://github.com/adafruit/circuitpython/pull/4110 fix https://github.com/adafruit/circuitpython/issues/4067 ? If so I'll link them and close it.

[slender iron]

nope

[idle owl]

Ok.

[slender iron]

it's a good question 🙂

[idle owl]

What does RTC support look like then?

Someone asked about it, and assumed it wasn't implemented because of that issue

[slender iron]

the PR enables python access to the RTC

the issue is about using the RTC internally for sleeping

[idle owl]

Ok

[jaunty juniper]

ah what about time.time() ?

[slender iron]

so, for users, they'll be able to use the RTC

[idle owl]

@jaunty juniper Support was merged 14 hours ago.

They may not be running the latest version.

[slender iron]

@jaunty juniper ideally it'd be hooked up but I'm not sure we have it connected yet

[manic glacier]

tannewt

[adafruit/circuitpython] New comment on issue #4067: Hook up RTC for sleeping and timekeeping

https://github.com/adafruit/circuitpython/pull/4110 Allows us to set the RTC. For sleeping, we need to modify https://github.com/adafruit/circuitpython/blob/main/ports/raspberrypi/supervisor/port.c#L164-L187 for sleep support.

[idle owl]

@slender iron Thanks!

[manic glacier]

tannewt

[adafruit/circuitpython] New comment on issue #4115: ESP32\-S2: hangs during startup if backlight pin is used in display init code

I suspect this is a bug in the ESP32-S2 PWM implementation used by display. I'd be ok merging it with the digitalinout code for now and leaving this issue open for someone to look into it.

[idle owl]

@tulip sleet Working on this CP Essentials page for Pins and Modules. Discussing import board. Is it worth including a pins.c reference? Or is that too in depth for this. I have it in my notes to discuss microcontroller.pin so those pin names will be referenced on this page somewhere.

[slender iron]

I think dir should be enough

[idle owl]

ok

[manic glacier]

tannewt

[adafruit/circuitpython] Pull request review submitted: #4049 Separate SSLSocket from Socket

This looks good to me! Thank you for separating these out!

tannewt

[adafruit/circuitpython] Issue closed: #4057 ESP32\-S2 "Out of sockets" in 6\.2\.0\-beta\.0

tannewt

[adafruit/circuitpython] Issue closed: #4101 SocketPool\.Socket\.\_\_hash\_\_\(\) is not immutable

tannewt

[adafruit/circuitpython] Pull request closed: #4049 Separate SSLSocket from Socket

[idle owl]

@slender iron Without checking pins.c, how can you tell what pin is which? For example, the QT Py has A0, A1, A2, A3, SDA, SCL, TX, RX, SCK, MISO, and MOSI on the silk. but 'A0', 'A1', 'A10', 'A2', 'A3', 'A6', 'A7', 'A8', 'A9', 'D0', 'D1', 'D10', 'D2', 'D3', 'D4', 'D5', 'D6', 'D7', 'D8', 'D9', 'I2C', 'MISO', 'MOSI', 'NEOPIXEL', 'NEOPIXEL_POWER', 'RX', 'SCK', 'SCL', 'SDA', 'SPI', 'TX', 'UART'] in board. How do I know which pin is D2 for example, without checking pins.c?

[slender iron]

try print(board.D2)

[idle owl]

board.D2 results.

[jaunty juniper]

I've been using that to map pins:

import microcontroller,board
for pin in dir(microcontroller.pin):
    if isinstance(getattr(microcontroller.pin, pin), microcontroller.Pin):
        pins = ["{:28s} ".format("microcontroller.pin."+pin)]
        for alias in dir(board):
            if getattr(board, alias) is getattr(microcontroller.pin, pin):
                pins.append("board.{}".format(alias))
        print(" ".join(pins))

microcontroller.pin.GPIO0     board.IO0
microcontroller.pin.GPIO1     board.IO1
microcontroller.pin.GPIO10    board.SD_CS
microcontroller.pin.GPIO11    board.IO11 board.SD_MOSI
microcontroller.pin.GPIO12    board.IO12 board.SD_CLK
microcontroller.pin.GPIO13    board.IO13 board.SD_MISO
[...]

[idle owl]

We have something like that for identifying SPI/UART/I2C combinations.

Looks similar except it tries to init pin pairs, and prints when it succeeds and fails.

@jaunty juniper That might come in handy though. Thanks for sharing it.

@slender iron That simply prints board.D2.

[slender iron]

@idle owl I don't know why there is a D2 in board if it isn't on the silkscreen

[idle owl]

@slender iron Because we seem to assign Dx names to everything even when it's not on the silk. 🤷‍♀️

[slender iron]

probably in the learn guide

[idle owl]

There's a great many more pins in board for QT Py than are on the silk.

Same with most boards. I2C etc aside.

[jaunty juniper]

ah I believe it's to give access to A pins as D pins ?

[slender iron]

the pins.c should have them listed so the top one matches the silk

[idle owl]

I guess that's my point. About mentioning pins.c. However, maybe I'll include Neradoc's script instead. We included scripts like that in the I2C/SPI/UART pages of this guide.

[slender iron]

the order should be reversed here: https://github.com/adafruit/circuitpython/blob/main/ports/atmel-samd/boards/qtpy_m0/pins.c#L10

[idle owl]

@jaunty juniper It is ok with you if I possibly include this script in a guide page?

[jaunty juniper]

like, they have D names so you know they can be used as Digital IO ?

[slender iron]

the first entry for a pin should be the one on silk

[jaunty juniper]

sure absolutely

[idle owl]

They can be used for digitalio if they are digital pins regardless of the pin name.

[slender iron]

and that'll change what prints when you do print(<pin>)

[idle owl]

Oh so that's a QT Py issue.

[slender iron]

the A and D naming is an unhelpful arduinoism

yes, and likely other boards too

[idle owl]

Oh........

ok

[stark ledge]

If I want to contribute a python library for CircuitPython - is there a canonical github repo for libraries that I should be sending PRs to ?

[idle owl]

I did print(board.A2) and it gave me board.D2. I see.

@stark ledge We have the CircuitPython Community bundle for community supported libraries, if that's what you're asking. https://github.com/adafruit/CircuitPython_Community_Bundle Are you saying you have a library you want to share? Or you want to contribute to a current library?

[stark ledge]

I have a new library that I wanted to add

Thanks ! that is exactly what I was looking for

[idle owl]

Excellent! We have a guide on contributing if you're interested. https://learn.adafruit.com/creating-and-sharing-a-circuitpython-library

[stark ledge]

Thanks that is perfect !

[idle owl]

@slender iron Alright, I think fixing all of the pins.c's right now isn't the best use of my time, so I think I'll include the script to find the pin map for now since the print method will be dubious at best if we've been that inconsistent with the naming. Thanks for the help.

[slender iron]

yup, that's ok with me

[jaunty juniper]

if you need to indicate the matching pin names, aren't they on the pinouts page of the guide ?

[idle owl]

@jaunty juniper In theory, yes. However, this page is going into a standalone guide. It's mean to explain what's going on with board and other built-in modules. In explaining board I want to enable folks to use the module to figure out what pin is what etc.

It'll end up mirrored into all of the board guides which means it will show up as being part of each board guide, but still. If you simply go through the guide without a board guide to go with it, you should be able to find your way around.

[jaunty juniper]

I see

[idle owl]

@slender iron If there's a microcontroller.pin.x without an associated board alias, does that mean it isn't broken out on the board? Or what's going on there.

[jaunty juniper]

I believe so

we can make a script that doesn't show the microcontroller pins

[idle owl]

😄 I was going to ask you if you could help me modify the script. @jaunty juniper

I want them included, but I think we shouldn't print out the ones without board aliases.

Actually. I see where you're going with it.

I think I want both versions. 🙂

[slender iron]

if it isn't in board I assume it's hard to get to

[idle owl]

Ok

[slender iron]

and I wouldn't recommend telling folks about them

[idle owl]

That's what I was thinking

[slender iron]

it's only useful for a new board

without a def

[idle owl]

Ok thanks

@jaunty juniper Ok, so I like your idea, but also like your current script. However, I'd like to modify it so it doesn't print any microcontroller.pin.x that doesn't have a board alias.

But I do like the idea of one that prints only the board aliases.

[crimson ferry]

the script doesn't print any microcontroller pins that don't have a board alias

[idle owl]

microcontroller.pin.PA03      board.A1 board.D1
microcontroller.pin.PA04      board.A2 board.D2
microcontroller.pin.PA05      board.A3 board.D3
microcontroller.pin.PA06      board.A6 board.D6 board.TX
microcontroller.pin.PA07      board.A7 board.D7 board.RX
microcontroller.pin.PA08
microcontroller.pin.PA09      board.A9 board.D9 board.MISO
microcontroller.pin.PA10      board.A10 board.D10 board.MOSI
microcontroller.pin.PA11      board.A8 board.D8 board.SCK
microcontroller.pin.PA15      board.NEOPIXEL_POWER
microcontroller.pin.PA16      board.D4 board.SDA
microcontroller.pin.PA17      board.D5 board.SCL
microcontroller.pin.PA18      board.NEOPIXEL
microcontroller.pin.PA19
microcontroller.pin.PA22
microcontroller.pin.PA23

08, 19, 22, 23?

[crimson ferry]

no, sorry, I take that back, but could be easily modified, I forgot I did that originally

[idle owl]

🙂 No worries

I could have been reading it wrong.

I can't figure out how to modify it. 😕 It's a bit beyond my skills.

[jaunty juniper]

https://gist.github.com/Neradoc/9ae8e456f0df15a3704eec72e7fa146c

[idle owl]

@jaunty juniper Nice!

[jaunty juniper]

it's missing a natural sort but well

[idle owl]

I don't think they need to be sorted necessarily

Or does sort do something I don't know about. Probably.

[crimson ferry]

oops, that's secret, I guess I shouldn't post the link o_O

[jaunty juniper]

oh it is ? my bad

[crimson ferry]

it's OK if I link to it?

[jaunty juniper]

sure

[crimson ferry]

thanks

[jaunty juniper]

I sorted them because I think it's easier to read alphabetically but the lack of natural sort irks me 😛 (natural sort is where D10 is after D9 instead of after D1)

[idle owl]

@jaunty juniper Ahhhh ok.

[crimson ferry]

dir does that too 😦

which I guess is the issue

[jaunty juniper]

natural sort is unfortunately rare in computers, as far as I know only the mac's Finder does it while on windows or linux you'll name your files "something001" etc. to keep the order correct

or with ls

and there is no way from board to know which is the silk screen name and make sure it's first or last in the line

[idle owl]

right.

Which is fine, really.

We're getting the info to folks in a pretty clean way.

They can figure out beyond that. 🙂

[idle owl]

@tulip sleet Are you around?

[tulip sleet]

@idle owl yup

[idle owl]

@tulip sleet Trying to explain the I2C/UART/SPI singletons. They eliminate the need for busio and simplify the code. What else do they do? Something about making it so you can use it multiple times without a pin in use error? Or is that not specific to the singleton.

[ionic elk]

@crimson ferry Hey, sorry I didn't get on this earlier, we had power knocked out by a snowstorm for most of the day

[tulip sleet]

the first time you call board.I2C() or similar, it will create the object. The times thereafter it returns the already-created object. Sometimes this is called "lazy initialization", though I am not sure if that is an intuitive term or not. It does not reserve the pins up front, so you could use those pins for something else if you never call board.I2C()

[ionic elk]

I'll try it out now

[idle owl]

@tulip sleet Hmm alright.

[tulip sleet]

or "created on demand"

[idle owl]

And instantiating it is not calling it? So like.... creating the sensor object doesn't call it until you use the sensor object?

[tulip sleet]

no, it's instantiated when you call it, but not before

the point is only that board.I2C() creates it on the first call, and returns the already-created object on subsequenct calls

[idle owl]

Ok

[tulip sleet]

maybe "lazy instantiation" is better

"instantiation" is ... not a term most people know

[idle owl]

Hmm yeah

[tulip sleet]

but if you can define it, great

[idle owl]

hehe

I'll try.

[tulip sleet]

"lazy creation"

I am not going to make you a grilled-cheese sandwich for lunch unless you ask for it. I am not going to make one in advance.

[idle owl]

Hah! 😂

[tulip sleet]

and when you ask for it again and again, I will give you the same one

(not exactly the best analogy, since it gets consumed0

[idle owl]

But I get the idea.

[manic glacier]

m4tk

[adafruit/circuitpython] New comment on issue #4115: ESP32\-S2: hangs during startup if backlight pin is used in display init code

@tannewt Thanks Scott for your fast reply.

In the meanwhile I have looked a little deeper. It seems that common_hal_pwmio_pwmout_construct() function couldn't find a non-reserved channel because the pwmout_reset() function is called to late.

pwmout_reset() is called by reset_port(). But in main.c the board_init() which does the display initialization is called before reset_port(). So some initializations are not done yet.

I changed the order quick&dirty for testing in mai...

[tulip sleet]

@idle owl re your proposed language:
I would not say board is just for pins: it's for any board-specific objects, most of which are pins. So I2C, etc. are default objects for those busio objects. You could also say "You can call board.I2C() as many times as you like, and it will always return the same object." just to emphasize the point.

those are not "pins"

[idle owl]

@tulip sleet I was wondering how to word that re: board. Thanks. Also I like that emphasis.

Right.

Exactly.

I almost asked earlier what they were, but remembered they are singletons, but didn't know how to broaden the explanation of board.

[tulip sleet]

stray + sign: Instantiating an object+

[idle owl]

Yeah I caught that already

hehe 🙂

• has been sneaking in a bunch lately.

@tulip sleet Thanks much. I am sure I will have more questions with this page.

[scarlet scroll]

hello, I have a question ... if I am writing a driver for an IC with circuit python and I need to write a non standard pulse e.g a low and delay for 88uS what are my options? time.delay() is obviously not very good, but would microcontroller.u_delay() is quite good but is it universal across all board implementations? Sorry if I am asking something that is easily found in docs

[tidal kiln]

@fossil gorge just fyi, it did end up being the matrix portal 🙂
https://learn.adafruit.com/matrix-portal-room-co2-monitor

@scarlet scroll "non standard" - so it's not something I2C, SPI, etc?

[scarlet scroll]

no DMX512 its basically standard serial but has an unusual make before break, make after break at the beginning.. its not that strict actually, I guess I am trying to understand with CP what people do when they have to do a funny toggle like that

[tidal kiln]

some interesting discussion here:
https://github.com/adafruit/circuitpython/issues/673

        time.sleep(88E-6)

that number looks familiar 🙂

[fossil gorge]

To true. Most people I knwo say "new up" (C#-ism?). I'm on team "Instantiate". Nice to see I'm in good apparent company!

[fossil gorge]

I thought that looked familar on Sunday night's Desk of Ladyada! That looks REALLY good! And I LOVE the smile/frown quick-look status.

[scarlet scroll]

thanks, helpful! hah funny, the first to code in this post points back to old micropython I wrote for it, full loop. https://github.com/adafruit/circuitpython/issues/673#issuecomment-442356216

[tulip sleet]

time.sleep() is never going to be that accurate, and even microcontroller.delay_us() is only a minimum, because there may be other things happening that delay things further. There is some RS485 support in the esp32-s2 and mimxrt10xx ports, but you need an external IC to get the electrical part right

[tidal kiln]

afaik, there's no simple way to do tight timing like that in user code

[manic glacier]

tannewt

[adafruit/circuitpython] New comment on issue #4117: Update pico\-sdk to 1\.0\.1

Thanks for trying @ajs256! I'll do it so I can get the correct timers for DMA.

[tidal kiln]

or, maybe, no way. at least with anything like a time. something approach.

[scarlet scroll]

Thank you, just wondering if there was a trick I was missing

[crimson ferry]

what about PulseOut?

[tulip sleet]

yes, PulseOut could be possible, if it can run fast enough

[tidal kiln]

cool. was thinking maybe that also.
https://circuitpython.readthedocs.io/en/6.0.x/shared-bindings/pulseio/index.html#pulseio.PulseOut

so you wouldn't bit bang with hand crafted timing delays. instead you'd preload the pulses and then send them.

[scarlet scroll]

Now that's a library I didn't spot, thanks! I'll test it out

[idle owl]

I went with it and tried to explain it.

@fossil gorge Speaking of you, did you get your PR sorted out with the licensing bits? Dylan is working through the few we missed. There was more than one. Glad we caught it, sorry it had to be an issue with your PR.

[fossil gorge]

I did - ish. I'm working around it locally by bypassing the pre-commit hook when committing to git (for now). Once the repo's been updated with the necessary licensing information, I will pull that down, and rebase my changes on top of that to ensure it's all good.

And no worries about it happening "on my PR". I've been on both the receiving and giving end of those "Oh yeah, this has to be done before that because of X" enough times. All part of the development experience 🙂

[idle owl]

Oh perfect, ok! Glad you're unblocked. Should be in the next day or two for it being updated. I'll have Dylan prioritise it so it's done tomorrow, because otherwise it will end up sitting until Monday.

[fossil gorge]

Sounds good. If Dylan can't get to it, like I said, I'm more than happy to do it for that repo as a separate PR 🙂

[crimson ferry]

Is there somewhere in the code where the overall memory map for CircuitPython is itemized... stack, heap, pystack, sleep memory, etc.?

[idle owl]

No worries, he replied to my prioritisation request. He said he'll do it first. 🙂

[fossil gorge]

Sounds good to me. Thanks for checking in and the update

[idle owl]

For all of our reviewers: I tested the GitHub CLI today. I run MacOS, so I have no idea if it works on other platforms (I imagine it must?). But it was super simple to install, and one command to pull down the code from a PR for testing. It clearly runs all of the git magic I usually have to do to check out a PR in the background, but it is definitely easier and it worked quite well. I highly recommend it if you simply want to grab some PR code for testing.

[manic glacier]

anecdata

[adafruit/circuitpython] New comment on issue #4103: ESP32\-S2: Configurable resources based on board RAM

Looks like PYSTACK is set to 1536 globally in py/circuitpy_mpconfig.h, but could be easily overridden at the port level (no existing overrides I could find) or board level (I think only TG-Watch does this).

In cases where there is sufficient external RAM for heap, what's the downside to using more internal RAM for IDF/sockets and pystack?

[manic glacier]

BiffoBear

[adafruit/circuitpython] Pull request opened: #4118 Clarified i2 c init error msg take 2

A fresh attempt at clarifying the I2C init error message, after I ran git train --wreck on the last one.

[tulip sleet]

Sounds like it's worth taking the time to learn it. I have been using the hub package for a while, which mainly lets you avoid having to type https://github.com/ most of the time, like hub clone adafruit/circuitpython, hub remote add somebody/circuitpython ,etc. gh does other stuff and I should learn it.

[solar whale]

@idle owl are there instructions for downloading and installing the CI stuff?

[idle owl]

@tulip sleet I think so, yes.

@solar whale I found the install command inside a PR. I opened an active PR in a browser, and in the upper right of the PR, there is an "Open With" button, and the resulting dropdown had the install command. It was MacOS specific, so I'm not sure if you're running something else, whether it shows you the command for your OS. In that same dropdown is the command to get the PR code.

I imagine there's a man page for it via commandline and I assume they've documented it somewhere, but I haven't looked further into it yet.

[solar whale]

Thanks. I’ll take a look.

[idle owl]

You're welcome!

[random oasis]

Does anyone know if the best way to get a Unique ID from the a board (specifically a m4) is microcontroller.cpu.uid?

[solar whale]

@random oasis I think that is the intent.

[tulip sleet]

yes, that is a unique id burned into each chip by the mfr

we might be using it in backwards order (there is some kind of canonical order), but that has not been a big deal: https://github.com/adafruit/circuitpython/issues/2087

[random oasis]

@solar whale @tulip sleet Do you know if there is a way to get the crc32 value of that id? (I've done this in the past with other languages) I'm trying to get a shorter, but almost always unique id, something that I can show on a display and someone could easily copy down.

[tulip sleet]

you could just xor some bytes with others

[random oasis]

xor?

[lone axle]

+1 for the GitHub CLI. I found it made things easier in the git workflow for me as well. I am on Linux, and it works fine there as well.

[fossil gorge]

As long as you remember to use gh repo fork ... if you don't have a fork. Otherwise, you have to manually setup the upstread with git remote ... commands. That's SO 2019

[still zephyr]

Hello, can I have some help I am trying to do my first PR, however the CI check did not pass. As far as I can see the error message is Process completed with exit code 1. in line 32. I want to understand what I am doing wrong. any help will be appreciated :). this is for https://github.com/adafruit/Adafruit_CircuitPython_Display_Text/pull/112

[manic glacier]

ajs256

[adafruit/circuitpython] Pull request review submitted: #4118 Clarified i2 c init error msg take 2

Looks okay to me (a grammar nerd), but some more experienced eyes should check first.

[idle owl]

@still zephyr It's not you. Something changes with a part of how our CI works and things are now failing.

[idle owl]

I think we only discovered it in the last couple of days. We're trying to figure out how best to fix it without creating a significant amount of future work for ourselves.

[still zephyr]

Oh thanks @idle owl. as a first timer, you always think is you 😆 . Thanks for the info.

[manic glacier]

ajs256

[adafruit/circuitpython] New comment on pull request #4118: Clarified i2 c init error msg take 2

Looks like the CI fail was for Trinket M0 in ja, that might need some shrinking.

[idle owl]

@still zephyr You're welcome. For your own reference, usually if the CI fails on something you did, it's forgetting to run black locally, or Pylint errors, and those are much clearer errors in the CI results. Black says it would reformat files, and Pylint gives you line numbers.

This one is a weird one, and isn't something you'd usually run into.

[manic glacier]

BiffoBear

[adafruit/circuitpython] New comment on pull request #4118: Clarified i2 c init error msg take 2

Looks like the CI fail was for Trinket M0 in ja, that might need some shrinking.

Thanks, this caused mayhem in my last attempt. @dhalbert has a fix, so I'll wait for his advice. I appreciate you taking the time to review my PR.

[still zephyr]

I see. question then, if I understand correctly I have to run black locally, maybe I miss that step, I was following the contribute to circuitpython and I did not see it.

[idle owl]

Oh! It's another guide. Maybe we didn't mirror it into the contributing guide. https://learn.adafruit.com/improve-your-code-with-pylint This covers Pylint and Black.

[manic glacier]

BiffoBear

[adafruit/circuitpython] Pull request closed: #4078 Clarified I2C init error msg

[umbral dagger]

Anyone have insight into issues with the FeatherSTM32F405 not playing nicely with the MPR121 STEMMA-AT breakout? Traceback (most recent call last): File "<stdin>", line 1, in <module> File "adafruit_mpr121.py", line 131, in __init__ File "adafruit_mpr121.py", line 180, in reset RuntimeError: Failed to find MPR121 in expected config state!

[still zephyr]

Thank you very much, I followed the guide and know the PR passed the checks. Thanks again and have a good night.

[idle owl]

@still zephyr You're entirely welcome! Well done!

[mental nexus]

ESP32-S2 question vs. pyportal: I'm running some bitmap manipulation (rotating some bitmaps) on the pyportal vs. the ESP32-S2 Saola, both running ParallelBus.

I'm surprised that the Pyportal is updating faster than the ESP32-S2. Is it possible there's some sleep-related issues that are affecting performance on the ESP32-S2?

Is there any way to ensure that the ESP32-S2 isn't sleeping?

[manic glacier]

hathach

[adafruit/circuitpython] New comment on issue #3986: Copying multiple larger files can hang the copy if autoreload is on

An alternative would be run all tasks while interrupts are off and then we reset before turning them back on.

This can be a bit tricky to do, since when background_callback_run_all() exit critical section to run task function, the tud_task() will then generate more usb isr to add to background queue. For this example, msc host will continuously bumping data fast enough, that will effectively causes MCU to actually have to complete the whole MSC copying before able to do the reload

...

[manic glacier]

dhalbert

[adafruit/circuitpython] New comment on pull request #4118: Clarified i2 c init error msg take 2

Just replace boards/trinket_m0/mpconfigboard.mk with this version:
https://github.com/adafruit/circuitpython/blob/595356d7fdd132c41c42a1441d41ade999b8f7b4/ports/atmel-samd/boards/trinket_m0/mpconfigboard.mk

hathach

[adafruit/circuitpython] New comment on issue #3986: Copying multiple larger files can hang the copy if autoreload is on

I make a quick hack to add usb background after the background_reset() for quick test, it does solve this issue and quickly complete the file copying without issue https://github.com/hathach/circuitpython/pull/1 . There is also a couple of minor code clean up as well.

[manic glacier]

ajs256

[adafruit/circuitpython] Pull request review dismissed: #4118 Clarified i2 c init error msg take 2

Looks okay to me (a grammar nerd), but some more experienced eyes should check first.

[manic glacier]

skieast

[adafruit/circuitpython] New comment on pull request #4077: Add support for Lilygo TTGO T8 ESP32\-S2

@tannewt updated and rebased. unfortunately I could not get the display_init code to work. I do have the code working in python. but, my C skills are just garbage.

@skieast I'll wait for your tinyuf2 PR. I suppose you already saw the PID for that.

I've created a PR for the tinyuf2 support.
https://github.com/adafruit/tinyuf2/pull/67

[idle wharf]

I did some research into options for dealing with the rev: in pre-commit.yaml.
It doesn't support mutable names (stable, latest) any longer and there is a wish to also not have to hardcode the versions or to manage it centrally.
Full notes on the tail end of this Issue's description. https://github.com/adafruit/cookiecutter-adafruit-circuitpython/issues/99

Summary:
There is one path where a custom local pre-commit hook is used to get a config file and then run pre-commit --config based on that file.
There is one path where where all repos are kept up to date by automating pushing changes to them. (links to options I found in the issue)
NOTE: Both of these options are suggestions from the pre-commit maintainer.

If anyone has additional suggestions, please add comments to that issue linked above.

[jaunty juniper]

oh the use of the screen for the bootloader on the ttgo is brilliant
when do we get that on the Clue (and others) ? 😄

[manic glacier]

BiffoBear

[adafruit/circuitpython] New comment on pull request #4118: Clarified i2 c init error msg take 2

Just replace boards/trinket_m0/mpconfigboard.mk with this version:
https://github.com/adafruit/circuitpython/blob/595356d7fdd132c41c42a1441d41ade999b8f7b4/ports/atmel-samd/boards/trinket_m0/mpconfigboard.mk

OK that's done. Thank you for your help and patience.

[manic glacier]

jfabernathy

[adafruit/circuitpython] Issue opened: #4119 sdcard demo missing baudrate attribute on Raspberry Pi Pico

After #4114 was closed, I tested a Pico with adafruit-circuitpython-raspberry_pi_pico-en_US-20210202-9140bfb.uf2 and the latest library bundle as 20210203. I modified the pin number in the demo and it still draws and error:
`Traceback (most recent call last):
File "code.py", line 23, in
File "adafruit_sdcard.py", line 101, in init
File "adafruit_sdcard.py", line 122, in _init_card
File "adafruit_sdcard.py", line 111, in _clock_card
AttributeError: 'SPIDevice' object has no ...

[manic glacier]

dhalbert

[adafruit/circuitpython] Pull request review submitted: #4118 Clarified i2 c init error msg take 2

Thanks for your patience as well, and for the original idea!

dhalbert

[adafruit/circuitpython] Pull request closed: #4118 Clarified i2 c init error msg take 2

[manic glacier]

jerryneedell

[adafruit/circuitpython] Issue opened: #4120 RP2040: charlcd demo fails

As reported by a Discord User, the accessing a Character LCD fails on a Raspberry Pi Pico

The same example runs normally on a metro-esp32s2

For this example I used SCL on GP1 SDA on GP0

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

Adafruit CircuitPython 6.2.0-beta.1 on 2021-01-28; Raspberry Pi Pico with rp2040
>>> 
>>> import charlcd_i2c_rgb_simpletest
Traceback (most recent call last):
  File "", line 1, in 
  File "charlcd_i2c_rgb_simpletest.py", line 18...

jerryneedell

[adafruit/circuitpython] New comment on issue #4120: RP2040: charlcd demo fails

Also ran against current tip of main -- same result

Adafruit CircuitPython 6.2.0-beta.1-68-gcd616f639 on 2021-02-03; Raspberry Pi Pico with rp2040
>>> import charlcd_i2c_rgb_simpletest
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "charlcd_i2c_rgb_simpletest.py", line 18, in <module>
  File "adafruit_character_lcd/character_lcd_rgb_i2c.py", line 75, in __init__
  File "adafruit_mcp230xx/mcp23017.py", line 52, in __init__
  File "adafruit_mcp230xx...

[manic glacier]

dhalbert

[adafruit/circuitpython] Issue opened: #4121 Add board\.I2C\(\), board\.SPI\(\), board\.UART\(\) for Raspberry Pi Pico board

It would be nice to have default board busio devices on the Raspberry Pi Pico. The question is which pins to pick.

In the Get Started with MicroPython on Raspberry Pi Pico book, the I2C examples use GPIO0 and GPIO1:

---

SPI uses GPIO2, 3, and 4:

---

![image](https://user-images.githubusercontent.com/2847802/1...

[manic glacier]

jerryneedell

[adafruit/circuitpython] New comment on issue #4121: Add board\.I2C\(\), board\.SPI\(\), board\.UART\(\) for Raspberry Pi Pico board

Not a big deal either way. Users have to learn to read the chart eventually ;-)
I like the first option. I expect there will be many folks following the book at first then trying CP.
Would you put the default UART on GP7 and GP8?

[idle owl]

@tulip sleet Around for a question about the issue you filed?

[tulip sleet]

on the phone, will ping you

[idle owl]

Ok thanks

[slender iron]

I don't know why it would be. Do you have wifi active while you are testing?

[manic glacier]

tannewt

[adafruit/circuitpython] Pull request review submitted: #4113 Translations update from Weblate

ladyada

[adafruit/circuitpython] New comment on issue #4121: Add board\.I2C\(\), board\.SPI\(\), board\.UART\(\) for Raspberry Pi Pico board

i think we try to matchy with the examples in the book, and add to the FAQ

[analog bridge]

@slender iron are you working on busio.UART for rp2040?

[slender iron]

I haven't started it yet

I've been thinking about how to do the dma/buffering

did you make progress on nvm?

[analog bridge]

yes... I'll PR tomorrow

[slender iron]

nice!

[manic glacier]

kattni

[adafruit/circuitpython] New comment on issue #4121: Add board\.I2C\(\), board\.SPI\(\), board\.UART\(\) for Raspberry Pi Pico board

How ubiquitous is the SDK within CircuitPython - which is to say, how likely is the typical user to be aware of the SDK info? Our goal is to make CircuitPython as simple as possible to use. As long as everything is documented, I think it's almost irrelevant what pins we use. That said, if we think the book is going to be/already is popular, we should follow those examples for simplicity. Otherwise, we should consider what the decision means in the long run - is the SDK going to be a longer-us...

[tulip sleet]

@idle owl did you state your q in the issue?

[idle owl]

@tulip sleet I did, figured that was a better place for it since you were busy and folks were replying already.

[manic glacier]

tannewt

[adafruit/circuitpython] New comment on pull request #4113: Translations update from Weblate

Network failure.

tannewt

[adafruit/circuitpython] Pull request closed: #4113 Translations update from Weblate

[idle owl]

@slender iron This is a comprehensive list of the built in circuitpython modules, yes? https://circuitpython.readthedocs.io/en/latest/shared-bindings/index.html#modules

[mental nexus]

@slender iron I took some time measurements and the nested for loop in my updated "blit" function (written in python) is taking about 75-100% longer on the ESP32-S2 than on the pyportal.

With the faster clock speed of the ESP32-S2 (240 MHz) versus the PyPortal (120 MHz) I suspected the S2 would outperform.

I was surprised by this, but there must be some differences that I don't understand. The inner loop is copying pixels from one bitmap into another. Should this memory access/writing be slower on the ESP32-S2?

[slender iron]

It's the best one I know of. There is a chance it won't have modules we inherit from MP

[tulip sleet]

it doesn't include those, see https://circuitpython.readthedocs.io/en/latest/docs/library/index.html

e.g. json, etc.

[idle owl]

Oh. Hmm.

But help("modules") wouldn't include those either?

[tulip sleet]

the support matrix is actually a good starting point

help("modules") does include those

[idle owl]

Ah ok

Then I'll link to both in the guide.

[tulip sleet]

help("modules") should be the same as the support matrix for any particular board

[slender iron]

memory access can definitely be slower due to external PSRAM and external flash

[idle owl]

@tulip sleet Fair enough, where the support matrix? I hadn't considered that.

[tulip sleet]

https://circuitpython.readthedocs.io/en/latest/shared-bindings/support_matrix.html#

[idle owl]

Ok. Yeah I'll need the other two links. I want to show everything available, then discuss how to find out what works on your board.

Thanks!

[tulip sleet]

in "Core Modules". We could probably hack the "Core Modules" .rst to point to json, etc., somehow, for now.

[mental nexus]

Ok, got it. Sounds like performance comparison between the two will not be so obvious to predict. Thanks for the feedback.

[idle owl]

@tulip sleet It would be ideal to have them all in one place.

[tulip sleet]

the real solution is to move all the modules to shared-bindings, but that's not done yet. I can see about fixing up the .rst's temporarily

[slender iron]

Yup, it's definitely more complex than clock speed. (Memory bandwidth is something the rp2040 should be good for.)

[mental nexus]

Also, I had assumed that the "python" overhead would be the limiting factor and not memory access time.

[tulip sleet]

@idle owl Do you have a UART example preference for default pins, since the book has no UART example?

[idle owl]

Makes sense. I'll tweak the wording in the guide for now and include both links I suppose. Bleh.

[idle owl]

There's no UART support yet I thought. There's no UART examples. We could follow the SDK in that one.

[slender iron]

I doubt it in this case because code gets cached as it's executed (rp2040 does this too) but RAM isn't cached afaik.

[idle owl]

As long as it doesn't clobber the other two anyway.

[tulip sleet]

unfortunately they would overlap with the book SPI pins

[idle owl]

Let me look at it and let you know what I think.

[manic glacier]

dhalbert

[adafruit/circuitpython] New comment on issue #4121: Add board\.I2C\(\), board\.SPI\(\), board\.UART\(\) for Raspberry Pi Pico board

I think the average CircuitPython user is not going to be looking at the MicroPython SDK document. The book is much easier to read and I can imagine it being a taking-off point. So using the first set makes sense. The main thing is to document it clearly.

[idle owl]

@tulip sleet SPI and I2C are numerical order, might as well arbitrarily go with 5/6..

@tulip sleet Actually

There's a ground pin between those two

[tulip sleet]

that overlaps with SPI

[idle owl]

Ok I'm missing something.

SPI in the book is 2,3,4

Are we doing more than one SPI?

[fringe trench]

ARGH!

I have been programming Python for TOO LONG, and it got SMARTER, and I DIDN'T NOTICE.

[tulip sleet]

it can't be GP5 and GP6 because they are not a UART pair

[idle owl]

@fringe trench 🎉

@tulip sleet I was going to change to 6/7 anyway.

Because ground pin.

So that coincidentally would work.

[tulip sleet]

has to be GP8, GP9 or high

[fringe trench]

You don't have to explicitly pass self to super().__init__ in Python 3, apparently.

[tulip sleet]

GP6/7 are not attached to a UART

[idle owl]

oh

I'm not reading GP#s I'm reading pin numbers.

Bleh

8/9 it is.

[tulip sleet]

i keep making the same mistake, it's a nightmare

having the silk labeled 1,2 is crazy

[idle owl]

@tulip sleet So frustrating.

[fringe trench]

@tulip sleet GP#, or super?

[idle owl]

Anyway, go with 8/9.

[tulip sleet]

GP 🙂

could you propose in the issue?

@fringe trench yes, that is nice in Python 3

the main trouble for a while was old examples on the web

[solar whale]

Does it mtter if it is UART0 or UART1 -- UART0 is on 12/13

[fringe trench]

You can tell how long it's been since I wrote class hierarchies...

goes to get a cane and slippers....

[tulip sleet]

@solar whale not really; MicroPython optionally has the REPL on UART0, but we are not proposing that

@idle owl I have a PR to add these in progress

[fringe trench]

Thank you for humoring me.

[solar whale]

I wish they did not use the term "SPI_TX/RX" confusing....

or is that the new standard replacing MOSI/MISO?

[idle owl]

Right on.

[manic glacier]

kattni

[adafruit/circuitpython] New comment on issue #4121: Add board\.I2C\(\), board\.SPI\(\), board\.UART\(\) for Raspberry Pi Pico board

As we do not have a UART example yet, please put UART on GP8/GP9. It's a UART pair, and avoids I2C and SPI suggested above.

[idle owl]

Done

[tulip sleet]

@solar whale SPI TX/RX is new to me. I know of MOSI/MISO, COPI/CIPO, SDI/SDO, but not TX/RX, which is ambiguous. https://www.oshwa.org/a-resolution-to-redefine-spi-signal-names

SDI/SDO are also ambiguous, as diagram shows.

[idle owl]

Ugh, it makes it like UART where they're backwards. I get that wrong now every time I'm wiring up SPI.

@tulip sleet I highly doubt you'll remember why I put this in my notes for this guide page, but I have "Special pin scl aka D2" as a bullet point, and I have no idea why. Any clue?

[tulip sleet]

the preview guide page from yesterday?

[idle owl]

Yeah

Supposed to be explaining board, import board, dir(board) and separately, built in modules.

I still need to put in an example of the ESP32S2 or Pico with different pin names, but I'm getting close to done.

but that was on my list, and I don't know why

[tidal kiln]

@solar whale i agree, about SPI_TX/RX. it's going to be nothing but wiring confusion. "try swapping wires" is going to have a new resurgence in trouble shooting land. 😦

[idle owl]

Welcome back to UART, I guess.

[tidal kiln]

yep

[tulip sleet]

@idle owl I don't know what the D2 note means, but is D2 also SCL or some board or another, like gemma or CPX or something?

[idle owl]

@tulip sleet D3 is SCL on CPX. I have no idea. I guess I'll skip whatever it was, and if Limor remembers, she'll catch it in review as missing.

Thanks for thinking about it with me.

This pin map script is super convenient. Thank you @crimson ferry and @jaunty juniper.

[tulip sleet]

maybe it was just a fictitious example of a pin with two names?

[idle owl]

Not sure why I would have said "special pin" though.

But yeah maybe.

And I covered that thoroughly.

[idle owl]

Hah, adafruit_bus_device messes up the layout now.

@tulip sleet Can you review the page for content accuracy? And copy edit while you're at it, but you do that naturally.

[tulip sleet]

@idle owl should I take over editing for copyediting?

or just tell you?

[idle owl]

@tulip sleet Take over for that, yeah. That would be great.

[blissful pollen]

Another issue to fix! 🙂

[idle owl]

Indeed!

[tulip sleet]

@idle owl, ok, made a number of edits for clarification

[idle owl]

Thank you so much!

[tulip sleet]

@idle owl Worth explaining: I removed "external" from "external library" because I thought what "external" meant wasn't that clear, and it might raise more questions.

[idle owl]

@tulip sleet I added that at the last minute because I thought it was needed. But yes. Good call.

[tulip sleet]

i also added "built-in" more places, as you'll notice

[idle owl]

Ok

[manic glacier]

StuartIanNaylor

[adafruit/circuitpython] New comment on issue #3948: Is there a plan to support esp32\-s3?

The esp32-s3 (S3) could be amazingly useful from KWS to image recognition but if Espressif try to shoehorn into Espressif only solutions of esp32-who & esp32-skainet its likely going to be a flop.
If we get anywhere near a complete port of tensorflow-lite then wow its got big potential.

[blissful pollen]

Apparently sleeping on a problem helps, realized the (dumb) mistake I was making getting ParallelBus on rp2040 going. The code is a mess still (and not optimized for the PIO totally) but works.

[gloomy shuttle]

@tidal kiln I created a new draft PR on the adafruit_nunchuk library to start the discussion about adding support for other accessories and if it should go here or be migrated to a community library. https://github.com/adafruit/Adafruit_CircuitPython_Nunchuk/pull/25

[tidal kiln]

anyone else have thoughts on bundle location for above?

[idle owl]

@tidal kiln If you feel it should go in the Community Bundle, then follow that.

[tidal kiln]

i'm fine with either and open to suggestions though. can see plus/minus either way.

[idle owl]

I understand, but I think it's going to end up falling to you to make the decision. I can see it working in both locations, but I think as Adafruit sells the nunchuk but not other things, it might be best in the community bundle for now. We can't easily support other controllers without one of us finding the hardware etc.

[tidal kiln]

that's why i lean toward community also

[idle owl]

Sounds like a plan, then.

[gloomy shuttle]

That makes sense to me.

[tidal kiln]

cool. thanks @gloomy shuttle . feel free to take whatever you want from the nunchuk lib.

[gloomy shuttle]

great thank you. I was hoping to keep the stuff from the nunchuk so there could be a single library that had everything and not have to point people back to the other in case.

[idle owl]

@gloomy shuttle I would include in whatever you submit to the community bundle, the refactor format you chose, though, to make it extensible. And if Adafruit starts carrying other controllers, we can easily port your code to nunchuk.

So make it a package with a base class like you have in the PR.

[gloomy shuttle]

Got it. Next question, should I start fresh using the cookie cutter and just copy code over

[idle owl]

That would be great.

[tidal kiln]

yah, having two libs to point people to isn't great (one of the minuses). but the implied hardware support is also important to consider.

[idle owl]

@gloomy shuttle You'll have to tweak some things post-cookiecutter to make it work for the Community Bundle. We have a PR in to make it so you can choose which bundle you're aiming for, but it's not merged yet.

[gloomy shuttle]

Okay, I think I ran into some of that when I first started this endeavor. I had to tweak some of the workflow/pre-commit stuff

[idle owl]

Yeah I think that covers it.

And setup.py as well I think... I'm not sure honestly.

But I know some things have to be tweaked.

That PR should be in today or Monday. (I'm the reviewer, and I'm only in Mon-Wed, so if I don't get a reply on my review today, it'll be next week.)

[gloomy shuttle]

Alright, thank you

[gloomy shuttle]

@tidal kiln any thoughts on what this library should be named? Should I name it CircuitPython_WiiChuck as I want to slightly mimic the Arduino WiiChuck library. https://github.com/madhephaestus/WiiChuck

[tidal kiln]

that seems fine

sounds better than CircuitPython_WiiMote_I2C_Accessory_Devices 🙂

[manic glacier]

m4tk

[adafruit/circuitpython] New comment on issue #4115: ESP32\-S2: hangs during startup if backlight pin is used in display init code

@tannewt Is it ok to call pwmout_reset() in the board_init()function to temporary solve the problem? if yes then I can try to create a PR (will be the very first 😯)

[idle owl]

Power failed. Sigh.

[manic glacier]

jfabernathy

[adafruit/circuitpython] New comment on issue #4119: sdcard demo missing baudrate attribute on Raspberry Pi Pico

I also get this error on ESP32-S2 Metro

tannewt

[adafruit/circuitpython] New comment on issue #4119: sdcard demo missing baudrate attribute on Raspberry Pi Pico

@jfabernathy Can you set it to this and try it? (In adafruit_sdcard.py)

self._spi.spi.configure(baudrate=250000)

Thanks for testing this!

jfabernathy

[adafruit/circuitpython] New comment on issue #4119: sdcard demo missing baudrate attribute on Raspberry Pi Pico

I'm going to need a little help the adafruit_sdcard I have is the one in the bundle which is a mpy version. Where to I get the .py version?

jerryneedell

[adafruit/circuitpython] New comment on issue #4119: sdcard demo missing baudrate attribute on Raspberry Pi Pico

Use the second button here : https://circuitpython.org/libraries

jerryneedell

[adafruit/circuitpython] New comment on issue #4119: sdcard demo missing baudrate attribute on Raspberry Pi Pico

Or go to the actual repository https://github.com/adafruit/Adafruit_CircuitPython_SD

[idle owl]

It's apparently a huge outage due to loss of power to the provincial grid. They're working to restore it, but ETA is ???

[idle wharf]

I'm going to send you what you need in a bit so we can get that merged.

[idle owl]

Lovely, thank you. Like you mentioned, I might catch something you wouldn't.

[manic glacier]

jfabernathy

[adafruit/circuitpython] New comment on issue #4119: sdcard demo missing baudrate attribute on Raspberry Pi Pico

So I got adafruit_sdcard.py and just put that in the CIRCUITPY/lib directory on the ESP32S2 and changed to self._spi.spi.configure(baudrate=250000)
Now I'm getting another error.

File "code.py", line 18, in <module>
 File "/lib/adafruit_sdcard.py", line 101, in __init__
 File "/lib/adafruit_sdcard.py", line 122, in _init_card
 File "/lib/adafruit_sdcard.py", line 112, in _clock_card
AttributeError: 'SPIDevice' object has no attribute 'chip_select'

[idle owl]

@idle wharf We're pretty confident at this point that the current answer is an Adabot patch is necessary and we'll pin everything. Correct?

[manic glacier]

tannewt

[adafruit/circuitpython] New comment on issue #4119: sdcard demo missing baudrate attribute on Raspberry Pi Pico

@jfabernathy try commenting line 112 out. The pin should already be high

[idle owl]

Hey I have power again!

[idle wharf]

Yes, I agree.

The alternative is writing or own hook (in its own repo) which we then use to get a global pre-commit.yaml.
Totally doable and would work with community repos too.

[idle owl]

Hmph. Perhaps in the future.

[idle owl]

Simply wanted to make sure I hadn't missed something before setting Dylan off on a mission.

[idle wharf]

No, you got it

[idle owl]

@fossil gorge https://github.com/adafruit/Adafruit_CircuitPython_ProgressBar/pull/24

[fossil gorge]

Wonderful news! Thank you 🙂

...once I can manage to get through this work day...

[manic glacier]

jfabernathy

[adafruit/circuitpython] New comment on issue #4119: sdcard demo missing baudrate attribute on Raspberry Pi Pico

with #112 commented out in adafruit_sdcard.py
line 18 in demo program, sdcard = adafruit_sdcard.SDCard(spi, cs)
still give this error

  File "code.py", line 18, in <module>
  File "/lib/adafruit_sdcard.py", line 101, in __init__
  File "/lib/adafruit_sdcard.py", line 166, in _init_card
AttributeError: 'SPIDevice' object has no attribute 'chip_select'

[idle owl]

@idle wharf Hmm, what do we do with setup.py in the community bundle version? Is it even possible to change it for the community bundle version? I'm not sure if it's templated or not.

Still includes # The project's main homepage. url="https://github.com/adafruit/Adafruit_CircuitPython_test_comm", # Author details author="Adafruit Industries", author_email="[e-mail]",

etc.

I'm not even sure the community bundle is deploying to PyPI, maybe we don't include it at all?

I commented on the PR.

[idle wharf]

I think there are many more things one could change in the community v. adafruit library. This probably isn't an exhaustive list, but I had thoughts on that here:
https://github.com/adafruit/cookiecutter-adafruit-circuitpython/issues/100

Those are my opinions though and I think someone like yourself should be the decider. It plays into how "CircuitPython" the literal string is used community libraries by default.
My thought would be yes, do it.

[idle owl]

@idle wharf Is it feasible to have it ask which bundle you're targeting first and give you different options based on that? Or is that not a thing.

[idle wharf]

Honestly, I'm not sure. I don't know much about cookiecutter (yet).

[idle owl]

Fair enough. If you want to look into that, consider it. If not, no worries. I'll reply in the issue you linked.

@lone axle I have another mission for you, should you choose to accept it. We're updating cookiecutter, and the Sharing a CircuitPython Library guide will need to be updated to go with it. Is this something you'd be willing to do?

[lone axle]

@idle owl yep I am up for that. I'm still planning to write a section on pre-commit for that Sharing a library guide as well. So definitely can make any other needed changes at the same time.

or around the same time at least, if not in the exact same sitting

[idle owl]

Ok great. I'll add you to the PR so you're aware of the timing.

@lone axle You can work with @idle wharf once the PR is merged to make sure the documentation is clear and accurate.

[crimson ferry]

this pylint report seems odd, it seems to want secrets before a native lib, and is calling wifi a "third party import"? https://github.com/adafruit/Adafruit_CircuitPython_Requests/pull/68/checks?check_run_id=1826084237

[idle owl]

@crimson ferry Sometimes it's weirdly picky. Does reordering the imports make it happy? Or does it continue to fail?

[crimson ferry]

I don't know, I was going to try to guide the author, but it makes no sense to me... do we blindly follow pylint's advice, or try to fix some underlying problem?

[idle owl]

@slender iron So the bus device built into CircuitPython now works? I confused myself when this sensor worked without me loading the bus_device library onto the board.

[crimson ferry]

I just haven't had much encounter with pylint before, so I don't know

[idle owl]

@crimson ferry Fixing an underlying problem would likely involve changing how Pylint works, and I don't think we're going to do that. If you don't agree with it, you can disable it in the code, but you should have a good reason for doing that. In this case, I would say follow it.

[slender iron]

yeah, is should. only sdcard seems unhappy about it

[crimson ferry]

@idle owl ok, cool, thank you

[idle owl]

Nice! Ok. Seems like this should have been a bigger deal. 🙂 One less thing to add to guides I suppose.

[idle owl]

@idle wharf FYI, the patch is in Adabot, but Dylan likely won't get to it until next week. He's prepared for the mission though, so in theory we'll have everything updated within a couple of weeks.

[tidal kiln]

@idle owl should this get archived?
https://github.com/adafruit/Adafruit_Python_LED_Backpack

[idle owl]

@tidal kiln Probably?

[tidal kiln]

i think this replaces it:
https://github.com/adafruit/Adafruit_CircuitPython_HT16K33

[idle owl]

Have you archived a lib before?

[tidal kiln]

yes

just came across someone using it forums. will figure out how they got there, in case a guide is still ref'ing it.

[idle owl]

Sounds like a plan.

[tidal kiln]

but i'm not finding anything

it has some reasonably recent issue/pr activity

but the last commit was 2018

[idle owl]

Hmm

I guess ping Limor.

If you're not sure about it. I know she wanted everything archived eventually.

[tidal kiln]

ok. yah. i probably will just to be sure there's not some odd reason it's being kept around.

[idle owl]

Sounds like a better plan.

Are you good with Arduino?

[tidal kiln]

i can blink an led 🙂

[idle owl]

If so I'm pinging you in the help-with channel because I'm not and it's not working and I don't even know where to begin troubleshooting.

lol.

Ok.

[tidal kiln]

i'll jump over there and take a look

[manic glacier]

tannewt

[adafruit/circuitpython] New comment on issue #3986: Copying multiple larger files can hang the copy if autoreload is on

I think your 2 idea is the best way to go about it for now. There isn't a single VM task.

We don't have a way to know which tasks to drop because we're stopping and those we can to keep doing. Perhaps we should add a critical flag or something.

tannewt

[adafruit/circuitpython] New comment on issue #4121: Add board\.I2C\(\), board\.SPI\(\), board\.UART\(\) for Raspberry Pi Pico board

My vote would be to follow the book.

However, I'm not a huge fan of adding default busses when the board doesn't document them itself. At some point its on the board maker to do the designation.

[fossil gorge]

Reading over the last few hour or 2 from #help-with-circuitpython, particularly the thread with coolmandude545, it made me realize something I'd noticed that felt disjointed, but I couldn't put my finger on it. But thanks to that conversation, I JUST did...

The issue I've had in the past, is that there's no obvious callout or indication which go from the code from a project's learn guide, to the documentation on ReadTheDocs. At least not on the projects I've been looking at or following

Just mentioning it since it occurred to me. Something that might be worth adding to guides, maybe on the code page/tab/section would be a listing of the libraries used, and links to the ReadTheDocs for them

[jaunty juniper]

yep, going from an example to the API is an involved process with much googling, and google often does not return the best version of readthedocs which can also lead to confusion

[fossil gorge]

Agreed for the RTD version. And I always forget to check that, and then need a moment to remember where the version selection is

I'm not familiar with the learning system backend, so I'm making some assumptions here... Something else that could be interesting from there, if the libraries for each guides are indexed, is a "other projects using library X". That way, it may make it easier to find different parts of the same library that can be used together.

I'm mostly thinking "out loud" about this, but figured I'd share the thought.

</ramble>

[manic glacier]

tannewt

[adafruit/circuitpython] New comment on issue #4115: ESP32\-S2: hangs during startup if backlight pin is used in display init code

@tannewt Is it ok to call pwmout_reset() in the board_init()function to temporary solve the problem? if yes then I can try to create a PR (will be the very first 😯)

Yup! (Usually the in use checks default to free.) If you find it gets reset you may need a never_reset call.

Thanks for debugging this!

tannewt

[adafruit/circuitpython] New comment on issue #4103: ESP32\-S2: Configurable resources based on board RAM

I think you'll find that the tradeoff is between IDF/sockets (which gets the remainder) and pystack.

[cedar moth]

As a new user of CP and the Adafruit products, I agree with Hugo that pointing users to the relevant ReadTheDocs from the guides would be very helpful. Finding a relevant guide is fairly obvious on the Adafruit website but, in my experience so far, going the next step to ReadTheDocs is not.

[manic glacier]

tannewt

[adafruit/circuitpython] New comment on issue #4120: RP2040: charlcd demo fails \- mcp230xx error

Looks like we could be getting an error from pico-sdk: https://github.com/adafruit/circuitpython/blob/main/ports/raspberrypi/common-hal/busio/I2C.c#L148-L157

[jaunty juniper]

then there is the arduous task of going through the inheritance tree

so I'm showing someone the doc to adafruit_display_text, and I'm puzzled because it doesn't have x and y properties listed, so I look at the source code, class Label(displayio.Group): , and now I'm looking at another RTD

[fossil gorge]

Ah, yeah. Inherited parts of a lib. Hmmmm

I wonder if Sphinx has a way to include the inherited portions as well. I know that I've seen it with GhostDoc and Doxygen, I think.

[slender iron]

yup, agreed we have good docs but they could be better organized

maybe we could parse example code and link each function/module out to readthedocs

[fossil gorge]

My thinking was, depending on how the guide's backend works, it might be easier to have the list of libraries as part of the guide's metadata

But like I said, that's making some assumptions about the backend

[slender iron]

learn is all custom so we could add it 🙂

I was hoping that circuitpython library RTD would have links to products

[fossil gorge]

That's the developer's answer right there! Not that I've ever said that! lol

[slender iron]

and guides do know products

I don't work on learn so I'm signing others up for work. 😛

[fossil gorge]

So yeah, that cross-referencing and links between libs, projects and docs would be awesome

[fossil gorge]

Well shoot. Here I was hoping to sneak something onto your "honey-do" list!

[slender iron]

it's already plenty long

[manic glacier]

tannewt

[adafruit/circuitpython] New comment on issue #4119: sdcard demo missing baudrate attribute on Raspberry Pi Pico

Ok, looks like sdcard takes in the chip select so it should just remember it.

[manic glacier]

hathach

[adafruit/circuitpython] New comment on issue #3986: Copying multiple larger files can hang the copy if autoreload is on

I think your 2 idea is the best way to go about it for now. There isn't a single VM task.

We don't have a way to know which tasks to drop because we're stopping and those we can to keep doing. Perhaps we should add a critical flag or something.

Ok, I will make an PR using this method

[manic glacier]

hathach

[adafruit/circuitpython] Pull request opened: #4122 Fix 3986 background reset remove existing usb callback

Issue

run_code.py decide to reload (AUTO_RELOAD) and run background_callback_reset() while background list is still queued up with usb event.

background_callback_run_all: 95:
callback_head = 200029A4
callback_tail = 200029A4
background_callback_run_all: 100: // background_callback_run_all() exit while there is queuing callbacks
run_code_py: 340:  // reload request == true
        usb_irq_handler: 103: // usb isr occurs -> more callback queued up
callback_head = 20002...

hathach

[adafruit/circuitpython] New review comment on pull request #4122: Fix 3986 background reset remove existing usb callback

I am not sure that we could just run usb_background() here, if it is then, we could remove the need to add this usb_background_schedule()

hathach

[adafruit/circuitpython] Pull request review submitted: #4122 Fix 3986 background reset remove existing usb callback

hathach

[adafruit/circuitpython] Pull request review submitted: #4122 Fix 3986 background reset remove existing usb callback

hathach

[adafruit/circuitpython] New review comment on pull request #4122: Fix 3986 background reset remove existing usb callback

this function name is the best I could come up, let me know if you want to rename it.

hathach

[adafruit/circuitpython] Pull request review submitted: #4122 Fix 3986 background reset remove existing usb callback

hathach

[adafruit/circuitpython] New review comment on pull request #4122: Fix 3986 background reset remove existing usb callback

the pointer should be volatile rather than the data it is pointed to. We actually can have it as

STATIC background_callback_t * volatile callback_head, * volatile callback_tail;

But I will leave it as both volatile just in case I misunderstand its purpose.

hathach

[adafruit/circuitpython] Pull request review submitted: #4122 Fix 3986 background reset remove existing usb callback

hathach

[adafruit/circuitpython] New review comment on pull request #4122: Fix 3986 background reset remove existing usb callback

this function seems to be redundant

[manic glacier]

kevinjwalters

[adafruit/circuitpython] New comment on issue #4034: Missing GP15 in board module for Raspberry Pi Pico

An interesting issue using GP15, Adafruit Forums: Code.py Startup Script Not Starting on Raspberry Pi Pico.

[manic glacier]

skieast

[adafruit/circuitpython] New comment on pull request #3349: Add Gravitech Cucumber R board definition

I have requested and received a PIDs from Espressif for the CucumberRIS board.
I have also created a PR for the UF2 Bootloader using the appropriate PID.

0x800C | GRAVITECH CUCUMBER RIS ESP32S2 - Arduino
0x800D | GRAVITECH CUCUMBER RIS ESP32S2 - CircuitPython
0x800E | GRAVITECH CUCUMBER RIS ESP32S2 - UF2 Bootloader

https://github.com/adafruit/tinyuf2/pull/69

[manic glacier]

gravitech-engineer

[adafruit/circuitpython] New comment on pull request #3349: Add Gravitech Cucumber R board definition

Thank you.

From: 'Bruce Segal' via Engineering [e-mail]
Sent: Thursday, February 4, 2021 8:33 PM
To: adafruit/circuitpython [e-mail]
Cc: gravitech-engineer [e-mail]; Mention [e-mail]
Subject: Re: [adafruit/circuitpython] Add Gravitech Cucumber R board definition (#3349)

I have requested and received a PIDs from Espressif for the CucumberRIS board.
I have also created a PR for the UF2 Bootloader usin...

[solar whale]

Just a heads up if anyone has time to see if they can reproduce: on a Raspberry Pi Pico , I have a 7 neopixel string (Jewel). If I set it up and write to indiviual pixels with ```
Adafruit CircuitPython 6.2.0-beta.1-80-gadaf43d6d on 2021-02-04; Raspberry Pi Pico with rp2040

import board,neopixel
pix = neopixel.NeoPixel(board.GP27,7)
pix[0] = 0xff0000
pix[0] = 0x0000ff

I'll dig further this afternoon and open an issue if no else gets to it first,

[manic glacier]

jepler

[adafruit/circuitpython] Pull request opened: #4123 raspberrypi: implement reset\_to\_bootloader

No testing performed.

[manic glacier]

m4tk

[adafruit/circuitpython] Pull request opened: #4124 Add display init code for Lilygo TTGO T8 ESP32\-S2

m4tk

[adafruit/circuitpython] New comment on pull request #4077: Add support for Lilygo TTGO T8 ESP32\-S2

@netroy I've created a PR for the display init code. #4124

[manic glacier]

netroy

[adafruit/circuitpython] New comment on pull request #4124: Add display init code for Lilygo TTGO T8 ESP32\-S2

i'll test and leave my comments later. thanks for doing this :)

[tidal kiln]

this is from Atmel Studio

anyone know how to get the equivalent of that Device signature number via JLink command line tools?

[tulip sleet]

@tidal kiln looking...

[tidal kiln]

thanks. i've been punching at it and nothing obvious.

[tulip sleet]

@tidal kiln

J-Link>mem32 0x41002018,1
41002018 = 10010305 

number above is from a Metro M0

[tidal kiln]

cool. thanks @tulip sleet

i was foolishly thinking there'd just be a command for it

vs. a low level read

[tulip sleet]

i hoped so too. It's a standard "CoreSight" thing that exists across all ARM chips.

this is a ROM location, not writable

[tidal kiln]

yep. thanks. was noting the R access. so really no way to shoot one self in foot and accidentally change that.

[tulip sleet]

by comparison, on a SAMD51J19:

J-Link>mem32 0x41002018,1
41002018 = 60060005 

[tidal kiln]

J-Link>mem32 0x41002018, 1
41002018 = 10010305 
J-Link>w4 0x41002018, 0x98bec2
Writing 0098BEC2 -> 41002018
J-Link>mem32 0x41002018, 1
41002018 = 10010305 
J-Link>

doesn't complain if you try. but doesn't do anything.

[tulip sleet]

I don't know what the problem we discussed earlier is, but changing the BOOTPROT fuses should not be related to the ID, so I suspect some kind of software bug

[tidal kiln]

most likely. was just wanting a second way to get that chip id value. and also verify there's really no way to accidentally mess it up.

do you know if there's a decoder ring for them somewhere?

[tulip sleet]

@tidal kiln the SAMD21 datasheet screenshot above is fairly informative, but the SAMD51 datasheet page is less so. I did some websearching for the various fields but found no master key.

[tidal kiln]

thanks. it must exist somewhere. atmel studio is comparing against something for "expceted" values.

[tulip sleet]

it looks like it's getting the wrong value when it tries to read it, but I don't understand how

[tidal kiln]

could there be a range of acceptable values? like to cover all known revisions. and it would need updating when new revisions come out.

[tulip sleet]

but in this case it's changing out from under someone to an unrecognized value

yes, there certainly would be a range because there are revision fields

[tidal kiln]

hmm. yah. and also 0x10010305 vs 0x1001031F would only differ in the DEVSEL bits

[manic glacier]

xorbit

[adafruit/circuitpython] Pull request opened: #4125 Add A3 pin to Raspberry Pi Pico

We want to be able to use this to measure VSYS.

hierophect

[adafruit/circuitpython] New comment on issue #4061: ESP32S2 time\.sleep\(60\) fails but time\.sleep\(15\) works

@jerryneedell can you retest this and see if it works for you now? My personal test works.

hierophect

[adafruit/circuitpython] New comment on issue #3994: Separate TLS implementation into ssl\.SSLSocket

Resolved by #4049

hierophect

[adafruit/circuitpython] Issue closed: #3994 Separate TLS implementation into ssl\.SSLSocket

[ionic elk]

@onyx hinge Could I just throw together a super quick PR to resolve #4023? I just noticed it while browsing the ESP32-S2 issues, it'd take me all of 10 seconds to add the missing claim-pin for each of the I2S construct pins.

[manic glacier]

jepler

[adafruit/circuitpython] New comment on pull request #4123: raspberrypi: implement reset, reset\_to\_bootloader

Updated and tested now :)

[onyx hinge]

@ionic elk that would be awesome of you!

[manic glacier]

FiriaCTO

[adafruit/circuitpython] Pull request opened: #4126 WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

I put a WEBUSB_README.md file in the top-level directory, but in brief - this PR incorporates the functionality from the tinyusb example program "WebUSB serial demo". When this option is enabled, websites like
https://adafruit.github.io/Adafruit_TinyUSB_Arduino/examples/webusb-serial/index.html can access a WebUSB serial console from the browser.

FiriaCTO

[adafruit/circuitpython] New comment on issue #3950: Support WEBUSB

The functionality of the tinyusb "WebUSB serial" demo has been incorporated into a webusb branch. It turns out you can have your CDC console and a WebUSB console at the same time, but depending on how many USB endpoints your chip supports, you might have to give up SOMETHING. For some ESP32-S2 work we are doing at Firia Labs, we have been trading USB HID for WebUSB at compile time. A PR has been submitted.

anecdata

[adafruit/circuitpython] New comment on issue #4061: ESP32S2 time\.sleep\(60\) fails but time\.sleep\(15\) works

Two issues I'm encountering post-4049:

1. HTTP doesn't work for me https://github.com/adafruit/Adafruit_CircuitPython_Requests/issues/63#issuecomment-771712906
2. Under certain circumstances, OSError: Failed SSL handshake occurs, particularly if there has been some delay since the previous Request [https://github.com/adafruit/Adafruit_CircuitPython_Requests/issues/63#issuecomment-771909006](https...

[manic glacier]

hierophect

[adafruit/circuitpython] New comment on issue #4061: ESP32S2 time\.sleep\(60\) fails but time\.sleep\(15\) works

@anecdata I'm looking into the HTTP thing now, sorry for the delay, I've been having power cutouts from the weather. As for the second point, are we sure that isn't defined behavior? I'm not an SSL expert but can an SSL session just persist indefinitely?

anecdata

[adafruit/circuitpython] New comment on issue #4061: ESP32S2 time\.sleep\(60\) fails but time\.sleep\(15\) works

The OSError: Failed SSL handshake seems to be correlated to no keep-alive, so either no Connection header, or Connection: close.

[ionic elk]

@crimson ferry @tulip sleet I have a weird python thing going on. When you do try: except SomeException: the docs imply it's supposed to look for that particular exception, but throw any uncaught exceptions that might happen, right? https://docs.python.org/3/tutorial/errors.html

[tulip sleet]

right

but SomeException could be a superclass of other exceptions

[ionic elk]

But I'm finding that this block:

try:
    print("REQ: send attempt")
    self._send_request(socket, host, method, path, headers, data, json)
except _SendFailed:
    print("REQ: send fail")

seems to just ignore whatever it doesn't catch

it just moves on to the rest of the program.

If I add another line:

except OSError:
    print("OSError")
else:
    print("Another Problem")

it'll land on these, but when they're gone, it just ignores them

[tulip sleet]

is there an enclosing try-catch that's catching the uncaught inside exception?

[ionic elk]

There is, and it should be, but it isn't.

it's just acting like there was no uncaught exception at all

[tulip sleet]

Adafruit CircuitPython 6.2.0-beta.0-43-g351a0e747-dirty on 2021-01-26; Adafruit Metro M0 Express with samd21g18
>>> try:
...     3/0
... except ZeroDivisionError:
...     print("caught it")
...     
...     
... 
caught it
>>> try:
...     3/0
... except ValueError:
...     print("caught it")
... 
Traceback (most recent call last):
  File "<stdin>", line 4, in <module>
  File "<stdin>", line 2, in <module>
ZeroDivisionError: division by zero
>>> 

Could you point to your code

[ionic elk]

Sure, I'm using @crimson ferry 's last example here: https://github.com/adafruit/Adafruit_CircuitPython_Requests/issues/63, and I'm modifying the lines described in this issue: https://github.com/adafruit/Adafruit_CircuitPython_Requests/issues/67

The same thing also happens with Jerry's code here: https://github.com/adafruit/circuitpython/issues/4061

[tulip sleet]

could you just upload the whole thing?

use the "+" on the left

[ionic elk]

That has messaging modifications in Requests

no secrets file

[crimson ferry]

the Requests mod I used to catch and show the trace on line 566 is:: except (_SendFailed, OSError) as e: sys.print_exception(e) ok = False

[ionic elk]

I'm finding that else: catches an error that is neither of those, but if you don't include it, it never gets caught

Don't know what it is yet

[tulip sleet]

@ionic elk you can do except Exception as e: and then print(e) so you can see what the uncaught one is

unfortunately you cannot do else as e:

i didn't see anything weird about the code you uploaded

[slender iron]

@tulip sleet are you planning on reviewing 4122?

thach's usb fix

[tulip sleet]

yes, I am testing it now

[slender iron]

ok! thank you!

[tulip sleet]

you were having a dialogue with him, so I though you were more enmeshed, but happy to t do it

[slender iron]

np, I was just trying to keep him moving

[manic glacier]

tannewt

[adafruit/circuitpython] New comment on issue #4034: Missing GP15 in board module for Raspberry Pi Pico

Here is a pending PR to restore the pad state: https://github.com/raspberrypi/pico-sdk/pull/80

tannewt

[adafruit/circuitpython] New review comment on pull request #4124: Add display init code for Lilygo TTGO T8 ESP32\-S2

Is this intentional?

tannewt

[adafruit/circuitpython] Pull request review submitted: #4124 Add display init code for Lilygo TTGO T8 ESP32\-S2

Code looks good! Will approve and merge after @netroy tests. Thanks!

ladyada

[adafruit/circuitpython] New comment on pull request #3349: Add Gravitech Cucumber R board definition

ok great @cotdp please update this PR :)

jerryneedell

[adafruit/circuitpython] New comment on issue #4061: ESP32S2 time\.sleep\(60\) fails but time\.sleep\(15\) works

@hierophect I just tried with tip of main and it still fails with time.sleep(60)-- do I need pull in a PR or update any libraries?

tannewt

[adafruit/circuitpython] Pull request review submitted: #4125 Add A3 pin to Raspberry Pi Pico

Ah! That's what it's used for! Do you have any reason to call it A3? We commonly have a VOLTAGE_MONITOR pin name on other boards. That'd make it clearer that the pin isn't broken out.

m4tk

[adafruit/circuitpython] Pull request review submitted: #4124 Add display init code for Lilygo TTGO T8 ESP32\-S2

m4tk

[adafruit/circuitpython] New review comment on pull request #4124: Add display init code for Lilygo TTGO T8 ESP32\-S2

Yes, GPIO4 is the LCD_MISO pin…that‘s why I removed it

hierophect

[adafruit/circuitpython] New comment on issue #4061: ESP32S2 time\.sleep\(60\) fails but time\.sleep\(15\) works

@jerryneedell you'd need to go into Requests and search for except _SendFailed and replace it with except (_SendFailed, OSError):

[slender iron]

@tulip sleet I assigned the webusb pr to you too. I think it fits into your usb thinking now

[manic glacier]

StarWitch

[adafruit/circuitpython] Issue opened: #4127 Socketpool \`socket\.sendto\(\)\` cannot send UDP data to ports above 255

Setup

• Microcontroller: UnexpectedMaker FeatherS2 (ESP32-S2) (fw: 0.2.1)
• CircuitPython: 6.1.0 & custom build from main
• Host computer: macOS 11.2
• Tools used: Wireshark

Description

It seems that port for socket.sendto() is bound to the length of an uint8_t, capping it out at 255, and then overflowing back to 0 when port numbers above that are specified.

I confirmed this by seeing seemingly random ports show up when I was trying to send data over port `9000...

hierophect

[adafruit/circuitpython] New comment on issue #4127: Socketpool \`socket\.sendto\(\)\` cannot send UDP data to ports above 255

Thanks for noticing this, definitely an oversight! Might have been an unintended carryover from another API.

dhalbert

[adafruit/circuitpython] New review comment on pull request #4122: Fix 3986 background reset remove existing usb callback

It's a good name!

dhalbert

[adafruit/circuitpython] New review comment on pull request #4122: Fix 3986 background reset remove existing usb callback

Tested while copying files: this works either way: usb_background() or usb_background_schedule(). I think they accomplish the same thing.

dhalbert

[adafruit/circuitpython] New review comment on pull request #4122: Fix 3986 background reset remove existing usb callback

Good catch. We can leave it as volatile both ways.

dhalbert

[adafruit/circuitpython] New review comment on pull request #4122: Fix 3986 background reset remove existing usb callback

I agree.

jerryneedell

[adafruit/circuitpython] New comment on issue #4061: ESP32S2 time\.sleep\(60\) fails but time\.sleep\(15\) works

@hierophect That appears to have fixed it for my initial example -- made it through 3 cycles so far

[tulip sleet]

@ionic elk did you figure out the disappearing exception? Is the whole program failing and restarting??

[ionic elk]

I'm still muddling through it, I think I'm missing something obvious

Going to take a quick lunch break and come back to it, clear my head

[manic glacier]

jerryneedell

[adafruit/circuitpython] New comment on issue #4061: ESP32S2 time\.sleep\(60\) fails but time\.sleep\(15\) works

@jerryneedell you'd need to go into Requests and search for except _SendFailed and replace it with except (_SendFailed, OSError):

Is this change being incorporated into Requests?

dhalbert

[adafruit/circuitpython] New comment on pull request #4075: turn on ticks during USB writes

Closed in favor of #4122.

dhalbert

[adafruit/circuitpython] Pull request closed: #4075 turn on ticks during USB writes

[manic glacier]

anecdata

[adafruit/circuitpython] New comment on issue #4061: ESP32S2 time\.sleep\(60\) fails but time\.sleep\(15\) works

I'll submit a PR

xorbit

[adafruit/circuitpython] New comment on pull request #4125: Add A3 pin to Raspberry Pi Pico

How about having both A3 and VOLTAGE_MONITOR names available? The RPi Pico docs seem to call it A3 so I would think it would be best to have this name available for the people coming from that side but VOLTAGE_MONITOR for the people who come from the Adafruit / CircuitPython side.

[manic glacier]

dhalbert

[adafruit/circuitpython] New comment on pull request #4126: WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

Thanks for doing this work. Before reviewing it in detail, I'd like to ask you if you also considered using the "Serial API", sometimes known as "WebSerial". This does not need its own descriptor on the device side, unlike WebUSB. Discussed in this closed issue: #605. #3950 is basically a duplicate of that.

Spec: https://web.dev/serial/
Status in Chrome: https://www.chromestatus.com/feature/6577673212002304

The Chrome demo does not...

dhalbert

[adafruit/circuitpython] New comment on issue #3950: Support WEBUSB

Note that another alternative is the "Serial SPI", sometimes known as "WebSerial". This was first mentioned in closed issue #605.

Spec: https://web.dev/serial/
Status in Chrome: https://www.chromestatus.com/feature/6577673212002304

The Chrome demo does not quite work with CircuitPython (I see repeated <DISCONNECTED> <CONNECTED>). But I think we could get it to work

[onyx hinge]

@tulip sleet thank you and Thach for working on that USB thing which I guess turns out to be my fault. I wonder whether the mistake is cancelling all background calls, vs just those that refer to objects on the gc heap ... since my real goal with that API was just to make sure those got cleaned away without directly tracking them all.

(I'm not sticking around to discuss it right now though, and I probably missed at least some of the discussion on the PRs for context)

[ionic elk]

@tulip sleet @crimson ferry Ok, I'm trying to sort out what exactly I should continue testing here

I had a misconception about how try except worked because I'm still not great at python and didn't read the docs right, so that's the disappearing exception out of the way.

It seems like the except (_SendFailed, OSError): solves at least part of the problem

[tulip sleet]

I don't think of it as a fault. Exactly what is going on in terms on USB tasks is tricky to understand. I thought there were things to do that were not triggered by an interrupt, and I still am not sure I understand it fully.

@ionic elk So what happened to the OSError exceptions?

[ionic elk]

Now I've got two things left: 1) @crimson ferry is HTTP still messed up for you? I don't really get what was going on with your code and 2) is the SSL connection timing out after about a minute and having to be replaced with a new socket expected behavior

[crimson ferry]

@ionic elk I just submitted a Requests PR with that fix, it masks the issue of Failed handshake, cause unknown.

[ionic elk]

@tulip sleet I thought that else was only supposed to catch exceptions. Because I am bad at reading.

So I thought it was catching an exception that was going away

[crimson ferry]

(2) the SSL handshake is failing for some unknown reason on certain circumstances, I don't know why. With the PR in Requests, it will get a new socket, but the failures will continue silently.

[ionic elk]

This is kind of a @slender iron question, maybe, since he wrote the SSL stuff

[crimson ferry]

(1) HTTP still doesn't work for me, the code block in the issue is complete, also fails the same way with the simpler/better requests.Sesion call.

(http = requests.Session(socket) instead of http = requests.Session(socket, ssl.create_default_context())) ...for insecure HTTP

[ionic elk]

My intuition is that SSL might just be timing out? As in, the handshake requires you to maintain the connection and not just go silent for minutes at a time. But I don't know

[crimson ferry]

well, except that it seems to happen when the TCP connection is closed, not when it's kept alive

[ionic elk]

Is your sense that the HTTP stuff goes down to the low level Socket layer? I haven't worked in the HTTP library at all

[slender iron]

¯_(ツ)_/¯

[crimson ferry]

I'm really not sure, I am by no means an expert

[ionic elk]

@crimson ferry I don't quite know what you mean by that last part

[crimson ferry]

regarding (2): it seems to happen when the HTTP/TCP connection is severed between Requests, rather than when it stays alive between requests

[ionic elk]

Me neither, I don't think anybody knows what's going on 😬

Really? I see it happening when it tries to fetch a socket, and can't reach it

REQ: attempting request
REQ: Got Socket 1
REQ: send attempt
ERROR: Failed SSL handshake
REC: try recvinto
REQ: Socket should close
REQ _close: attempt to close
REQ _close: closed
REQ _close: deleted
REQ _close: end
REQ: Socket closed
REQ: Got Socket 2
REQ: send attempt
REC: try recvinto
REQ: break
REQ:

[crimson ferry]

in any case, re: (2), user code is happy with the Requests patch, we may just need to keep in mind that something deep down may be lurking

I'm not sure how to interpret that log, perhaps IDF is doing something counterinituitive when it doesn't get a Connection: keep-alive, or maybe it's not even aware of that level?

[ionic elk]

^Here's what my internal messaging indicates: it fetches an existing socket with _get_socket, and then attempts _send_request, which fails with an SSL handshake exception. Because this is now caught, it automatically closes that socket, because it's invalid, and opens a new one to try again

[crimson ferry]

right, so user code is happy, but we really don't know why the handshake failure

it's reproducible, circumstances seems constrained (not just any old HTTPS request)

[ionic elk]

My suspicion is that it's a timeout - I'm not an SSL expert, but maybe it's like USB, where if you don't maintain it, it'll die on you.

Because it works with short periods, but after a minute, the socket goes "stale" and can't be used anymore.

I have no idea if that's true, and tbh I'm not really sure where to start on researching it 😕

[crimson ferry]

maybe it's because IDF/Requests is not explicitly doing any behavior based on presence (and value) or absence of an HTTP Connection header ??

[manic glacier]

FiriaCTO

[adafruit/circuitpython] New comment on pull request #4126: WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

Some consideration was given to "Web Serial", and in fact if we had not been able to add support for "WebUSB Serial", the newer "Web Serial" was our backup plan. The reason we had such a high level of interest in "WebUSB serial" is that we have existing products already shipping that use it.

The reason we went with WebUSB originally is that the project/product line predates the addition of "Web Serial" to Chrome.

On a related note, we are also going to try to make use of WEB_DFU on this...

[crimson ferry]

Is IDF purely TCP level? If so, then it's something that Requests would handle.

[ionic elk]

That I don't know anything about, I'm just coming from the Sockets perspective. I don't even understand why we're using a module called HTTP for HTTPS.

[crimson ferry]

But I'm fine with masking it for now, and we see if something crops up later.

[ionic elk]

We now have two different versions of Socket - the regular Socket is TCP only, and SSLSocket, which wraps a Socket, uses MBED TLS under the hood.

I still would like to help with this HTTP error but I don't properly understand what it is or whether it's something I should be helping with on the C end

[crimson ferry]

and those TCP sockets can carry any higher-layer protocol presumably... so I'm thinking this is a limitation of current Requests

[ionic elk]

By carry, you mean you add additional application layer code on top of them?

[crimson ferry]

@ionic elk I'd propose we call (2) done, leaving only the insecure HTTP case.

right, like HTTP, HTTPS, etc

[ionic elk]

ok

[crimson ferry]

@slender iron, if you're following, does this make sense?

[ionic elk]

SSLSocket simply looks like Socket with a more restricted set of functions

When it wraps an existing Socket, it creates an mbed TLS object. Connect calls esp_tls_conn_new_sync, which maybe is the Handshake part? And it does mention a timeout. But I don't know where that code came from. @slender iron did you write that stuff from scratch or were you working from an example?

[manic glacier]

hierophect

[adafruit/circuitpython] New comment on issue #4061: ESP32S2 time\.sleep\(60\) fails but time\.sleep\(15\) works

I'm closing this issue now as @jerryneedell's remaining issue is in the Requests library and is continued in this issue: https://github.com/adafruit/Adafruit_CircuitPython_Requests/issues/67

hierophect

[adafruit/circuitpython] Issue closed: #4061 ESP32S2 time\.sleep\(60\) fails but time\.sleep\(15\) works

[ionic elk]

@crimson ferry Ok when I try to run your sketch I immediately get:

Traceback (most recent call last):
  File "code.py", line 23, in <module>
  File "/lib/adafruit_requests.py", line 613, in get
  File "/lib/adafruit_requests.py", line 567, in request
  File "/lib/adafruit_requests.py", line 435, in _get_socket
RuntimeError: Sending request failed
Traceback (most recent call last):
  File "code.py", line 39, in <module>
NameError: name 'response' is not defined

[crimson ferry]

HTTP? right, that's what I get, I can't do an insecure HTTP at all anymore

[ionic elk]

can you just quickly reiterate what this is supposed to be doing?

[crimson ferry]

just doing an HTTP GET on a URL

but not TLS

[ionic elk]

Ah, ok

Yes, what's probably going on here is that the Requests library automatically attempts to create a TLS socket no matter what?

so it doesn't even have an option for regular HTTP

[crimson ferry]

I don't know, Requests often baffles me, it used to work when sockets were combined

[ionic elk]

I don't know which Socket call is used for an insecure HTTP get, I can try and look that up.

Right, that was when we had both plain TCP and TLS sockets side by side under the hood and you didn't know what calls were using what. So you could be totally unencrypted for some stuff when you thought you were.

I thought that Requests might break a bit given the separation but didn't know to what extent.

[crimson ferry]

How should Requests get a non-TLS socket?

[ionic elk]

I think this simply deserves a new issue for Requests

I don't know, is Requests based off a Cpython module?

[crimson ferry]

wouldn't it just be by leaving off the wrap/default-ssl-context?

[ionic elk]

^oh, mechanically, yes. You just don't wrap it, and it'll stay TCP

[crimson ferry]

Scott may know

or Brent

[ionic elk]

@prime flower would you be able to chime in here?

[crimson ferry]

So, in theory, http = requests.Session(socket) instead of http = requests.Session(socket, ssl.create_default_context()) should bypass TLS socket and get a regular one? That still produces the same exception trace.

[prime flower]

I'm pretty sure requests is based off https://requests.readthedocs.io/en/master/

circuitpython-requests is "requests-like", or I;'ve seen it described like that

[ionic elk]

Yes, if you detected the lack of an ssl context, you could just make a regular socket

[crimson ferry]

"you" being the core circuitpython Socket code?

or the library, or me the user?

[ionic elk]

The library

[crimson ferry]

but how does the library do that? what's the API?

I thought it was as above

[ionic elk]

I'm saying maybe it should work as you said above? That http = requests.Session(socket) should make an non-TLS socket? I just meant to say I can see how that would work.

[crimson ferry]

do we need to be passing another parameter to socketpool? or socket?

[ionic elk]

I would say let's check the source code of @prime flower's link and see how they do it. Cpython shares the Socket and SSLSocket distinction, it's what my code is based off - we were non-conforming before the split.

[crimson ferry]

my question is: what API should the library use to get a non-TLS socket? https://circuitpython.readthedocs.io/en/latest/shared-bindings/socketpool/index.html

not invoking ssl for one

[ionic elk]

I mean to say that I don't know, and I'm the wrong person to ask, I'm just looking this up as I go. My ballpark is C drivers, not the HTTP layer, sorry!

[prime flower]

@crimson ferry this is also of interest to me in the app layer for minimqtt

I do

https://github.com/adafruit/Adafruit_CircuitPython_MiniMQTT/blob/master/adafruit_minimqtt/adafruit_minimqtt.py#L284

which wraps the socket if context has been provided, otherwise i dont wrap the socket.

[crimson ferry]

I really don't have a clear idea of where the issue is

[prime flower]

Is that incorrect? is it always making a SSL socket?

[crimson ferry]

I don't know, HTTPS is working great now, but no HTTP

[ionic elk]

@prime flower if you don't wrap the Socket it isn't a SSLSocket.

[prime flower]

@crimson ferry this is on esp32s2 right?

[ionic elk]

there's no SSL in a regular <Socket> object anymore.

[crimson ferry]

@prime flower yes, ESP32-S2

[prime flower]

ok,

[crimson ferry]

when we don't wrap the socket, getting a socket for HTTP fails https://github.com/adafruit/Adafruit_CircuitPython_Requests/issues/63#issuecomment-771712906

(see the edit at the bottom of the comment to use: http = requests.Session(socket))

[prime flower]

@crimson ferry and wrapping a socket for a HTTP url succeeds?

[crimson ferry]

this also doesn't work for an HTTP URL: http = requests.Session(socket, ssl.create_default_context())

I can't figure out a way to get HTTP URLs to work

[manic glacier]

hierophect

[adafruit/circuitpython] Pull request opened: #4128 ESP32S2: Add I2S pin claiming

Resolves #4023

[prime flower]

@crimson ferry I'm a bit tied up today with work but I'm curious if this also effects insecure (non SSL) mqtt sockets in minimqtt..

[crimson ferry]

that would be an interesting test

[prime flower]

I could try isolating the socket creation code from the library and connecting to netcat

[jaunty juniper]

I'm having problems with non SSL sockets too in my websockets code

[ionic elk]

@tulip sleet quick style question for you. Is there a preference to use mp_uint_t in the common_hal layer rather than a simple uint or unsigned int?

Both socket modules are a bit all over the place right now

[tulip sleet]

if you are going to be passing it back to shared-bindings, you should use mp_uint_t, since that signifies the type that is used internally for non-Python ints. But uint and unsigned int are both ambiguous. Perhaps it should be uint32_t for those places?

[manic glacier]

FiriaCTO

[adafruit/circuitpython] New comment on pull request #4126: WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

Re-reading and actually clicking on links, I see the "Chrome Demo" you referenced was in regards to the "Web serial" API.

I thought you were saying you had trouble with the "WebUSB Serial" demo at https://adafruit.github.io/Adafruit_TinyUSB_Arduino/examples/webusb-serial/index.html

[tulip sleet]

I almost never use bare int, since its size can be in question

[ionic elk]

@prime flower @crimson ferry my raw Socket examples are still working on Main but if anything comes up that seems like it might be a low level problem feel free to tag me

@tulip sleet yeah uint32_t is what I typically used for all my common hal stuf

[tulip sleet]

you would use an mp_uint_t if it was going to be converted to or from a python object

[manic glacier]

chbarg

[adafruit/circuitpython] New comment on issue #4007: NRF52: Support for Zigbee Protocol

I am sort of considering pivoting to the XBee3 which already runs MicroPython out of the box and can be bought as dev boards or modules. All the (nasty) low-level ZigBee code abstraction is already dealt with. At this point I do not see the point in all the hassle with trying to reinvent the wheel on nrf52.

As far as I know, the XBees do not have the Zigbee cluster communicati...

[ionic elk]

Well, these are parameters being passed in from Python into the common_hal layer

It's stuff like what should bool common_hal_socketpool_socket_connect(socketpool_socket_obj_t* self, const char* host, size_t hostlen, mp_uint_t port); use

[tulip sleet]

yes, that makes sense there

[ionic elk]

But that's inconsistent with the other common hal modules like Busio

which also directly take in parameters but rarely use an mp_ prefixed typedef

[crimson ferry]

It should "just work", but does anything in here look extra or inconsistent with our LWIP implementation? https://github.com/espressif/esp-idf/blob/6e776946d01ec0d081d09000c36d23ec1d318c06/examples/protocols/http_request/main/http_request_example_main.c

[ionic elk]

Should I also change things like the size_t used for hostlen, or the const char* for host?

[tulip sleet]

can you give me some examples, e.g. I see this:

The .u_int is declared as a mp_int_t (here u_ does not mean unsigned)

[ionic elk]

@prime flower I don't know quite how I'd interpret that. Our stuff is implemented based on the LWIP examples provided by the ESP-IDF and I tested it with basic server and client examples on my host machine over my local network

[tulip sleet]

was the size_t converted from a python int? I don't think so. That's ok

[ionic elk]

@tulip sleet so, in shared-bindings/busio/I2C.h and most other shared-bindings header files, we use exclusively types like uint32_t

in shared-bindings/socketpool/Socket.h we had kind of a hodgepodge and since that code started out with a basis from Micropython I was wondering if there was a reason for why the older functions were using mp_uint_t

[tulip sleet]

it doesn't matter -that- much; the busio code was written a long time ago and might not be completely consistent

[ionic elk]

my preference is to do uint32_t for everything, but I wanted to make sure that's the right call

I don't like the micropython types because it seems inconsistent with the rest of the common_hal code which looks more like traditional embedded C

[tulip sleet]

well, be careful you validate the incoming python arguments as being >=0 before passing on to common-hal

[ionic elk]

wait mp_uint_t should still be unsigned, right?

[tulip sleet]

we might eventually use a 64-bit processor

it is, but you have to make sure any incoming python arg is >= 0 before casting it to unsigned, otherwise you will end up with a very large value.

As a (bad) example, the frequency value in I2C() should be non-negatvie, but that currently isn't checked. It should be.

[ionic elk]

Well, that seems like maybe a different project, going through all the inputs and making sure they're sanitized.

A good project though, I'll make a note of it.

[tulip sleet]

If you grep for mp_int_t and mp_uint_t in ports/*/common-hal, you will find a lot of uses. It's a mistake that the busio routines pass an mp_int_t on one side and receive a uint32_t on the other.

[ionic elk]

For now I can make sure the inputs to Sockets are sanitized though.

oh, hmm.

[tulip sleet]

yeah, complain about negative timeouts, etc.

[ionic elk]

I guess this is an opportunity for a wider cleanup.

I still don't like that mp_ prefixed stuff is in common_hal though, that doesn't seem right to me. I don't like the ambiguity of it in a scenario where we're doing bit-level math

At the very least, I'd like to make it consistent. Do we have mp_ versions of size_t, char, uint8_t, etc?

[manic glacier]

dhalbert

[adafruit/circuitpython] New comment on pull request #4126: WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

Re-reading and actually clicking on links, I see the "Chrome Demo" you referenced was in regards to the "Web serial" API.

Yes, sorry, I explicated the link.

I tried the one you mentioned, and it won't let me "Add a port". I can access the same CDC UART via putty (with WebUSB running too).

I wonder if some sort of "experimental flag" needs to be enabled in Chrome...

I think I must have enabled it ages ago, when it first appeared behind a flag in Chrome 80-ish. It is suppose...

FiriaCTO

[adafruit/circuitpython] New comment on pull request #4126: WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

I wonder if some sort of "experimental flag" needs to be enabled in Chrome...
Yep... I happen to be running Chrome 88.0.4324.150 so I had to enable "Experimental Web Platform features" and "Use system serial port enumerator", but I did get it to finally bring up the REPL in the Chrome demo.

lwerner-lshigh

[adafruit/circuitpython] New comment on issue #4045: func \.\_\_ code\_\_ ?

I think i have a solution... not as elegant but ill just provide the "metadata" by hand as a property.

class Program(object):

    '''to be filled out in programs to let us know the input types they want'''
    props = {}

    def exec():
        '''to be implemented in programs'''
        pass

    def run(self):
        inputParams = {}
        for param in self.props:
            inputParams[param] = self.props[param](
                    input(f"{param}: ")
     ...

lwerner-lshigh

[adafruit/circuitpython] Issue closed: #4045 func \.\_\_ code\_\_ ?

[tulip sleet]

mp_int_t and mp_uint_t are meant to designate integers that were converted from a Python object. If you want to cast them to uint32_t, that's fine. I think it's just clearer if that's explicit. If it makes sense for the the common-hal routine to take a uint32_t, you could do that.

See grep mp_int_t shared-bindings/*/*.h for instance

lots of mp_int_t's being passed in; do the same for mp_uint_t as well, there are fewer

[prime flower]

@ionic elk Yeah this is some functionality I didn't see mirrored over from cpython, "If a non-zero value is given, subsequent socket operations will raise a timeout exception"

[tulip sleet]

mp_uint_t is weird, because it implies it's been vetted as >= 0. mp_int_t is more typical

[slender iron]

I'm not following along. Just back from lunch. I think you two are past my knowledge anyway.

[slender iron]

Examples

[manic glacier]

FiriaCTO

[adafruit/circuitpython] New comment on pull request #4126: WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

I still think it is worth having the option to build with WebUSB support. Since that standard has been defined longer, it may make it into other web browsers sooner.

[jaunty juniper]

is this the same problem you are investigating ?
this code https://gist.github.com/Neradoc/469fe2918083c2a4bc3d584e481079c2

• ends properly on 6.1
• gives me OSError: [Errno 9] EBADF at connect() on latest S3
• ends properly on latest with SSL = True
  (FeatherS2)

[manic glacier]

dhalbert

[adafruit/circuitpython] New comment on pull request #4126: WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

Chrome is somewhat of an outlier in trying to implement WebUSB and the Serial API; you can see in the links above that Firefox and other browsers have decided not to support it in the interest of safety.

I am fine with the basic idea of this PR: we'll just leave it turned off by default. Unfortunately at the moment my own PR has a bunch of additions and cleanups to gen_usb_descriptor.py, and I'm also changing how to specify which USB and USB HID devices to provide, so there's a lot of c...

[ionic elk]

@prime flower what do you mean? My two most recent PRs have improved the timeout system from what it originally was - you can pass in 0 for no timeout, which will result in EAGAIN for calls that would block, actual timeouts which will result in ETIMEOUT, or None, which is passed into the common_hal as (uint)-1 and is specifically caught to make sure it never times out.

@jaunty juniper you're closing the socket before you run connect()

that's why you get a EBADF

@slender iron do you have a preference on where we should sanitize and cast mp_uint_t into things like uint32_t and uint8_t? Common hal or shared bindings?

[slender iron]

I'm not particular about it

[manic glacier]

dhalbert

[adafruit/circuitpython] Pull request review submitted: #4126 WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

This is fine to include turned off by default. I think we will tend toward the Serial API to avoid needing new descriptors, but this is valuable.

Just two minor things.

dhalbert

[adafruit/circuitpython] New review comment on pull request #4126: WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

www.circuitpython.org redirects to circuitpython.org, so leave off the www.. But if this is an actual URL, shouldn't it have https://?

dhalbert

[adafruit/circuitpython] New review comment on pull request #4126: WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

// These next two hardcoded descriptors were pulled from the usb_descriptor.c file

[jaunty juniper]

I don't though

[ionic elk]

@jaunty juniper wait nevermind yeah I read your code wrong, my bad

thought it was with, I'm off my game today

[manic glacier]

dhalbert

[adafruit/circuitpython] Pull request review submitted: #4126 WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

dhalbert

[adafruit/circuitpython] New review comment on pull request #4126: WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

Same comment as above re www. and https:// prefix

FiriaCTO

[adafruit/circuitpython] Pull request review submitted: #4126 WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

FiriaCTO

[adafruit/circuitpython] New review comment on pull request #4126: WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

No objection to leaving off the www. The C struct that uses this has a different field that controls the prefix, look for bScheme.

FiriaCTO

[adafruit/circuitpython] Pull request review submitted: #4126 WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

FiriaCTO

[adafruit/circuitpython] New review comment on pull request #4126: WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

Same answer. OK to drop the www and the scheme is hardcoded to https:// via the bScheme field

dearmash

[adafruit/circuitpython] New comment on pull request #4088: Add support for watchdog timer to atmel\-samd

@dhalbert Just checking in to see what your thoughts are.

One refactor I can do, just rolling back up a loop.

The other with timers, I'm understanding a little bit, there's some library to help manage resource conflicts iiuc. I'm just curious if it's just a question of migrating to use an existing library, or if the existing library doesn't exist / is under-featured and will need more work.

FiriaCTO

[adafruit/circuitpython] Pull request review submitted: #4126 WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

FiriaCTO

[adafruit/circuitpython] New review comment on pull request #4126: WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

I just updated the supervisor.mk file.

dhalbert

[adafruit/circuitpython] New comment on pull request #4088: Add support for watchdog timer to atmel\-samd

@dearmash Sorry, I haven't had time to look at this in more detail. I think you will be able to use an existing clock generator, since we have several at different rates, but we have to check the clock speed and the divisors. The library is very low-level and assumes little infrastructure, and we have to fit that into the current clock infrastructure, etc.

FiriaCTO

[adafruit/circuitpython] Pull request review submitted: #4126 WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

FiriaCTO

[adafruit/circuitpython] New review comment on pull request #4126: WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

URL changed to omit www. in gen_usb_descriptor.py

dhalbert

[adafruit/circuitpython] Pull request review submitted: #4126 WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

change www.circuitpython.org to circuitpython.org in gen_usb_descriptor.py

dhalbert

[adafruit/circuitpython] New review comment on pull request #4126: WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

OK, go ahead and change this.

dhalbert

[adafruit/circuitpython] New comment on pull request #4126: WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

Ah, we crossed!

[blissful pollen]

@slender iron Two quick questions about ParallelBus on RP2040. 1) I have the PIO program set to 60Mhz any higher seems to not work, I couldn't find anything in the spec docs for the display of the maximum, so guessed at that.

2. I found an issue in the StateMachine code I believe. I can fix it in the PR I will submit soon or could make it its own if you prefer?

[slender iron]

does parallelbus take a frequency? seems like it should

[manic glacier]

FiriaCTO

[adafruit/circuitpython] New comment on pull request #4126: WebUSB serial support \(compile\-time option, currently defaulted to OFF\)

Sorry for the "cross"... can you fix it from your side?

[slender iron]

nice! you can fix it 🙂

[blissful pollen]

No frequency in the constructor i'll double check

[slender iron]

probably wise to add it since we'll start seeing mcus faster than displays can handle

[blissful pollen]

RP2040 already is 🙂 I'll take a look at adding that as well

[slender iron]

thanks!

[ionic elk]

@tulip sleet so there are no helpers to use in shared-bindings already that check whether an incoming argument is negative and throw an exception/cast it if so?