#circuitpython-dev

1 messages · Page 385 of 1

idle owl
#

Yeah seems like an issue that could be resolved.

#

Nice!

gilded cradle
#

Tried hot glue at first, but that kept coming off eventually

idle owl
#

I got one of the Pimoroni rainbow cases, and a STEMMA SHIM, so now I can test breakouts without that monster acrylic piece with the cobbler and a breadboard on it.

gilded cradle
#

Nice

idle owl
#

I thought maybe the SHIM would be blocked by the case

#

but weirdly, it's not.

#

The gap for the header is wide enough for it to fit perfectly

gilded cradle
#

Awesome

idle owl
#

Now I need to actually set it up. 😄

#

@tulip sleet I don't see a pullup on this one anywhere. But as we found on the previous one, I might not be finding it. Hallowing M4 Express

Screen_Shot_2021-08-19_at_12.11.17.png
orchid basinBOT
idle owl
#

I can't remember how to do raw tone code.

#

Trying to test the Hallowing M4 and getting nothing.

#

Code is running, not failing. But no sound.

#

I'm sure it's me not doing this properly.

tulip sleet
idle owl
#

Oh.

#

I was trying to use sinewaves and RawSample.

#

Hmm the speaker is an invalid PWM pin.

#

@tulip sleet I can't use PWM and the built in speaker?

#

I found our raw audio code. Finally. 🤦🏻‍♀️

#

But it's still not working.

tulip sleet
#

oh, sorry, yeah. I wouldn't bother with sinewaves. Just use a wav file

#

use audioio.AudioOut

idle owl
#

oh fair enough

crimson ferry
idle owl
crimson ferry
idle owl
#

Good call.

#

This was a waste of time.

#

If that's accurate, then the naming is consistent.

#

MONSTERM4SK doesn't even have one built in, it's a Molex PicoBlade connector to add one if desired.

tulip sleet
idle owl
#

I gave up and moved on.

#

Not finding pullups on MonsterM4sk or PyBadge either.

#

PyBadger lib says enable=True to turn it on tho.

tulip sleet
#

If you see /SD on a schematic, then it will be enable=True

idle owl
#

Ohhhh ok

#

I had it backwards in my head.

tulip sleet
#

not-SD means active-low shutdown (pull down to shutdown), so that means active high enable

idle owl
#

I am fuzzy on active-high vs. active-low. I mean, I understand it means pulling it to 3.3v or GND to enable it, but it doesn't gel in my head as quickly as I'd like it to when you make that statement. I have to think about it more than I'd like to admit before it starts to make sense, and I'm not even sure I'm right.

#

OH wait.

#

The pullup doesn't make the true or false different, it makes it default to on instead of off.

#

🤦🏻‍♀️ sigh.

idle owl
tulip sleet
#

I think you got it: active-high means when it's 3.3V (or whatever voltage), i.e. True, then the thing that's being named is "on" or "enabled" or whatever

manic glacierBOT
slender iron
#

@tulip sleet welp, the usb issue was a problem we had fixed on other ports already

#

sleeping when we have work to do

#

I'm tweaking it to check for any background task

#

may help audio too

tulip sleet
#

well, you have the Beagle now for whatever future highspeed debugging you want to do (e.g. RPi)

slender iron
#

yup yup 😄

#

it was helpful

tulip sleet
#

i could do beta.0 soon to include this. I am trying a SAMD audio fix but it may not work

slender iron
#

it might be this same issue

#

it's a race between running background tasks and then deciding to sleep

tulip sleet
#

I am making sure that the buffer that finished playing is the one that will be filled next, same as what I did on RP2040, but that used two DMA's, this uses one.

#

but that may help too

slender iron
#

👍

#

worth checking

manic glacierBOT
idle owl
#

@tulip sleet Ok. I've gone through everything with a SPEAKER_ENABLE pin, and they all seem to be /SD. So I don't think anything needs INVERTED.

tulip sleet
#

that is great, so the changeover is confined to macropad

idle owl
#

Yeah. Based on what Limor said yesterday, I think it might be ok to remove the SHUTDOWN pin now.

#

Thoughts?

#

I'd have to update the MacroPad library, but that's all I'm aware of.

tulip sleet
#

maybe it's in guides?

idle owl
#

It might be in the Pinouts page of the main guide, but that would be it.

#

Let me check.

tulip sleet
#

I did a google search over learn, and that seems to be it. It was called speaker_shutdown in MagTag arduino

idle owl
#

I just found that.

manic glacierBOT
idle owl
#

Yeah the guide has the SHUTDOWN pin.

#

So I'd update that too.

#

@tulip sleet CPB has SHUTDOWN as well.

tulip sleet
#

in CPy?

idle owl
#

Checking.

tulip sleet
#

{ MP_ROM_QSTR(MP_QSTR_SPEAKER_ENABLE), MP_ROM_PTR(&pin_P1_04) },

idle owl
#

No.

#

Only in the guide is it referenced.

#

Should be updated in the guide apparently.

#

Ok, so I can update the MacroPad library and the guide.

#

Library should wait for another release though.

tulip sleet
#

a quick PR and I'll get it into beta.0. Scott found the i.MX problem so that's a good incentive for beta.0

idle owl
#

Oh good ok.

#

I will remove the pin and push, and then PR.

tulip sleet
#

i can update the CPB guide immediately if you want

idle owl
#

Sure that would be great.

tulip sleet
#

done

manic glacierBOT
idle owl
#

That was a journey. 😄 Learned some things along the way, so it was worth it, I suppose.

#

Back to trying to templatise the MP3 page.

manic glacierBOT
idle owl
#

Woops. Borked my MacroPad. CIRCUITPY is NO NAME. sigh.

manic glacierBOT
#
jepler
[adafruit/circuitpython] Issue opened: #5183 Convert all circuitpython modules to use MP\_REGISTER\_MODULE

This can reduce the amount of boilerplate required. Instead of introducing a macro such as QRIO_MODULE dependent on CIRCUITPY_QRIO and then adding an expansion of it in MICROPY_PORT_BUILTIN_MODULES_STRONG_LINKS, one simply writes in the bindings __init__.c file:

MP_REGISTER_MODULE(MP_QSTR_qrio, qrio_module, CIRCUITPY_QRIO);

where the first item is the module name as a qstr, the second item is the module's globals, and the third is the preprocessor macro that expands to (1...

onyx hinge
pastel panther
#

hey @tulip sleet , I've got a few Adafruit_CircuitPython_BLE and/or Adafruit_Blinka_bleio questions

tulip sleet
#

sure

pastel panther
#

In short, have you seen any issues with doing a scan on M1 macs?

tulip sleet
#

let me look in bleak

pastel panther
#

ok; this works fine for me (old x86 macbook) but crashes hard on an M1 mac mini:
for adv in self._ble.start_scan(ProvideServicesAdvertisement):

tulip sleet
#

user did not give Terminal bluetooth permission

pastel panther
#

hmm 👀

tulip sleet
#

What I did to fix it was to add my Terminal app (in my case iTerm2.app) into System Preferences -> Security & Privacy -> Bluetooth and I just drag and dropped the app there.

onyx hinge
#

wow crashing your whole program seems a bit draconian

tulip sleet
pastel panther
#

will do, thanks

manic glacierBOT
onyx hinge
#

we could enable keypad in qtpy haxpress, ... there are 80 bytes left in fr, the fullest build with it off. oof.

#

we could do a qtpy_keebs build with busio off and keypad on, but we don't really want to go down the road of having a bunch of build profiles for otherwise identical hardware

manic glacierBOT
manic glacierBOT
manic glacierBOT
#
dhalbert
[adafruit/circuitpython] New review comment on pull request #5123: Check for duplicate pins in rows and columns

Instead of calling three routines here, suppose validate_no_duplicate_pins_2() called validate_no_duplicate_pins() on each of the lists it was passed. I can't think of a case where it's OK to have duplicates between the lists. So then a single call does all the validation. It doesn't make a difference in this PR but if validate_no_duplicate_pins_2() were called elsewhere it would save some calls.

manic glacierBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
#
waiweng83
[adafruit/circuitpython] Issue opened: #5186 Print function doesn't print the GPIO state correctly\.

CircuitPython version

Adafruit CircuitPython 7.0.0-alpha.6-103-gf9f106b58 on 2021-08-20; Cytron Maker Pi RP2040 with rp2040

Code/REPL

import board
import digitalio
import time

button = digitalio.DigitalInOut(board.GP20)
button.direction = digitalio.Direction.INPUT

led = digitalio.DigitalInOut(board.GP0)
led.direction = digitalio.Direction.OUTPUT

while True:
    led.value = not button.value
    print(button.value)
    
    #time.sleep(0.1)
`...
manic glacierBOT
manic glacierBOT
manic glacierBOT
#
waiweng83
[adafruit/circuitpython] New comment on issue #5186: Print function doesn't print the GPIO state correctly\.

I don't see you set a pull direction in your code ?
Does the board have a pull up ?

button = digitalio.DigitalInOut(board.GP20)
button.direction = digitalio.Direction.INPUT
button.pull = digitalio.Pull.UP

Alternatively you can use this:

button = digitalio.DigitalInOut(board.GP20)
button.switch_to_input(digitalio.Pull.UP)

It doesn't matter as there is external pull up on my board (Maker Pi RP2040).
Also the LED does...

#
dhalbert
[adafruit/circuitpython] New comment on issue #5186: Print function doesn't print the GPIO state correctly\.

I think you may be seeing switch bounce. You are reading the button state twice, and the second time may be different than the first, due to bouncing. The delay does debouncing for you.

Also, the LED value is the opposite of the button state. You did not say what you expected, just that you thought it was wrong.

Try this:

while True:
    button_now = button.value
    led.value = not button_now
    print(button_now)
    
    #time.sleep(0.1)
#
waiweng83
[adafruit/circuitpython] New comment on issue #5186: Print function doesn't print the GPIO state correctly\.

I think you may be seeing switch bounce. You are reading the button state twice, and the second time may be different than the first, due to bouncing. The delay does debouncing for you.

Also, the LED value is the opposite of the button state. You did not say what you expected, just that you thought it was wrong.

Try this:

while True:
    button_now = button.value
    led.value = not button_now
    print(button_now)
    
    #time.sleep(0.1)

...

bleak matrix
#

Hi all 👋

In board there are aliases for pins that have certain functions e.g. SD_CS. That got me wondering why the pins for I2S audio are absent - it would be jolly nice! Has this been discussed before or does anybody have any thoughts regarding this? Cheers

onyx hinge
#

@bleak matrix good question! do you have a particular board in mind, or are you just asking generally? If the hardware has an I2S DAC "attached", or there's just one set of pins, or the pins are called out that way on the silk or reference, I think it'd be a good idea to add. If there are a bunch of sets of pins and none of them are any more "official" than the others, then probably not. But there is a lot of grey/gray area, and what I said before is just my first gut reaction to the general case.

bleak matrix
#

It was actually for my SAMD51 board. From my understanding the M4 boards have fixed pins for I2S. Today, I had to go though the documentation - fortunately Adafruit have good docs 😀
However, for the RP2040 PIOs then aliases don't make sense as you say.

idle owl
#

new pi who dis

#

installs all the things

lone axle
#

@tulip sleet I am working though the last few items on the breaking changes list. One of them is Removal of gamepad from HID library I don't think that library is using the core module gamepad but it does have a gamepad.py file for emulating video game controllers. Is the idea with this one that we want to rename that to something else?

jaunty juniper
#

the existing gamepad controller descriptor was not very compatible with much things, so the support for it was moved from the library to the examples, so you would rather make your own descriptor (and the example could be renamed to avoid clash where the unrelated builtin gamepad module is left in)
the breaking change is there is no longer adafruit_hid.gamepad

manic glacierBOT
lone axle
#

For renaming it I would propose GameController or even VideoGameController but definitely open to input from others.

manic glacierBOT
#
NathanY3G
[adafruit/circuitpython] Issue opened: #5187 Add aliases for I2S audio pins to board module

In the board module there are aliases for pins that have certain functions e.g. SD_CS. It would be convenient if similar aliases existed for the I2S audio pins. This would make it unnecessary to add conditional logic similar to the following:

board_name = os.uname().machine

if "Grand Central M4 Express" in board_name:
    i2s = audiobusio.I2SOut(board.D14, board.D33, board.D32)
elif # etc.

I believe a number of boards have fixed I2S pin assignments including:

  1. Ad...
bleak matrix
#

BTW Thanks too all who have worked on the audio in Circuit Python. Being able to listen to tunes via C.P. brings a smile to my face 😃

idle owl
#

TFW you have to test a CO2 sensor, and end up finding out your air quality is terrible.

manic glacierBOT
#
waiweng83
[adafruit/circuitpython] New comment on issue #5186: Print function doesn't print the GPIO state correctly\.

I don't have the Maker board, but I was unable to replicate this on an QT Py RP2040 running absolute newest from S3 (7.0.0-alpha.6-121-ga98363348 on 2021-08-20), mirroring the Cytron Maker Pi RP2040 button circuit here: https://drive.google.com/file/d/1Zp8GYO8x7ThObB1G8RIZx2YdqrXtdUc0/view
The LED and print() both reflected the correct state, regardless of delay.

This is strange. I'm able to reproduce the problem running QT Py RP2040 firmware on Maker Pi RP2040.
I'm also able to re...

ember iris
idle owl
manic glacierBOT
ember iris
idle owl
#

I ran the code, and I'm thinking, that seems a little high, but what do I know, so I googled acceptable levels, and oi.

ember iris
#

I'm afraid to do that with the pm2.5 levels of my sensors..

idle owl
#

Yeah I feel like I should hook one of those back up and run it again now that I've gotten the CO2 levels in here.

ember iris
#

It's a nice one to have running, but I'm way more mindful of cooking now. I made a steak a while back and maxed out the sensor levels due to the smoke.. I've also used it to see how long ago I put something in the oven--it's pretty useful for that since I'm really forgetful

#

circuit python made the data logging a breeze

idle owl
#

For sure.

manic glacierBOT
tulip sleet
# lone axle <@!329766224093249548> I am working though the last few items on the breaking ch...

Hi - "gamepad" is the official HID name for controllers with those page/usage-page numbers, so I think we should keep that. I'll comment more in https://github.com/adafruit/Adafruit_CircuitPython_HID/issues/71

GitHub
Fix gamepad examples · Issue #71 · adafruit/Adafruit_CircuitPython_...

gamepad.py is no longer in the base library, because the gamepad device was removed from CircuitPython 7.0.0. Rename examples/gamepad.py to examples/hid_gamepad.py or similar, because gamepad.py ov...

cunning crypt
#

Hitting ctrl+c in the REPL should cease whatever's running, right?

tulip sleet
#

yes, a few bugs w/r/t to that were just fixed

blissful pollen
cunning crypt
idle owl
manic glacierBOT
onyx hinge
#

@tulip sleet my gut says to just reject low i2c baud rates with a new test, rather than try to modify the sercom's clock dynamically (asf4 is not ready for that), what do you think? Not much below 100kHz is possible

tulip sleet
#

is this true on SAMD21 also

#

i guess it has a lower clock rate

onyx hinge
#

Otherwise if I change the clock in asf4_conf it lowers spi top rate too

#

samd21 has the same max divisor 255

tulip sleet
#

i think it's fine not to support rates below 100kHz; we were just doing it for debugging

onyx hinge
#

But the base clock is 12MHz maybe?

tulip sleet
#

I see that the hpl routine sets BAUDLOW to zero

#

yah, maybe. We have much more use for higher SPI than lower I2C

onyx hinge
#

That's "same as baud", I think they mean high and low times

#

Or driven and undriven

tulip sleet
#

you can always pass a low clock to bitbangio.I2C()

onyx hinge
#

It's not 8 more bits of divider

tulip sleet
#

i forget that stuff

onyx hinge
#

I just learned much of that

# 
``` we'd know if this arithmetic brought in double precision math, right?
tulip sleet
#

I would see if SPI has a lower bound check, if not I would say don't bother to complain

#

we would see aebi_dmul or whatever it's called, and we would have seen that a long time ago

onyx hinge
#

for SPI we limit the divisor to 255, so it would be capped instead of giving anomalous higher rates

#

samd_peripherals_spi_baudrate_to_baud_reg_value in peripherals

#

so I could also implement that: lower cap at 100kHz

tulip sleet
#

it would be better to always round down, so as not to overtax a peripheral that would be marginal at higher speed

#

I saw the 238kHz (?)comment go by in email, but I can't find it

manic glacierBOT
#
lesamouraipourpre
[adafruit/circuitpython] Issue opened: #5189 Remove \`set\_vertical\_scroll\` from displayio

As far as I can tell the set_vertical_scroll parameter in the Display constructor is not used.

grep -r set_vertical_scroll *
shared-bindings/displayio/Display.c://|     def __init__(self, display_bus: _DisplayBus, init_sequence: ReadableBuffer, *, width: int, height: int, colstart: int = 0, rowstart: int = 0, rotation: int = 0, color_depth: int = 16, grayscale: bool = False, pixels_in_byte_share_row: bool = True, bytes_per_cell: int = 1, reverse_pixels_in_byte: bool = False, se...
onyx hinge
tulip sleet
#

so that is an overflow problem or something, it needs to confine it to eight bits

onyx hinge
#

24MHz/97 instead of /2401 or something

#

yes but asf4 blithely throws away the upper bits, rather than returning an error

#

so the easiest fix is just a hardcoded check so we don't have to touch asf4

tulip sleet
#

ugh, sure

#

I was just reading some DMA threads in the Atmel forums: "the ASF4 example is just wrong....", etc.

onyx hinge
#

asf4_conf/samd21/peripheral_clk_config.h:#define CONF_GCLK_SERCOM1_CORE_FREQUENCY 48000000 limits seem the same for samd21, assuming this is right

#

asf4_conf/samd51/peripheral_clk_config.h:#define CONF_GCLK_SERCOM1_CORE_FREQUENCY 48000000

tulip sleet
#

so we have lived with this forever, we were only trying lower I2C freqs for debugging, and that was highly misleading, unfortunately

onyx hinge
#

That is kinda what it looks like to me

#

I wonder, does arduino make it work

manic glacierBOT
slender iron
#

@tulip sleet sorry I missed that I had broken the unix build on the background check pr. I know you wanted to do a beta

tulip sleet
#

I don't think I'll do it today, maybe monday. I am on a good trail on SAMD51 audio

#

np

manic glacierBOT
#
dhalbert
[adafruit/circuitpython] New comment on issue #5187: Add aliases for I2S audio pins to board module

The functionally labelled pins usually correspond to either silkscreen markings or hardwired connections (e.g. NEOPIXEL, SD-CS). We don't tend to assign function names to pins such as the Grand Central pins above when you'd have to look up the pins anyway. Though there are boards where only certain pins can be used for certain functions, like the I2S examples above, there are other boards where the choice of pins is much wider. For instance, the nRF chips can do I2S on any pins.

The SA...

idle wharf
#

Hello Circupers (we need a logo...)

I've been looking at this PR and testing Circup and I am stumped as to why pimoroni_circuitpython_ltr559 needs to be special cased.

https://github.com/adafruit/circup/pull/120

Anyone have time to take a look and can help me understand ?

jaunty juniper
#

the pimoroni library doesn't follow the rule: the repository is named pimoroni_circuitpython_ltr559 but the module is named pimoroni_circuitpython_ltr559.py when it should be called pimoroni_ltr559.py

idle wharf
#

OK so the rule that is coded into cookiecutter is wrong then

jaunty juniper
#

I don't know, did they make it with cookie cutter ? does cookie cutter enforce the name for the library ?

#

I mean, yes they had to use cookie cutter, but does it chose the name of the library ?

idle wharf
#

yes... I'll make one, one sec

manic glacierBOT
#
lesamouraipourpre
[adafruit/circuitpython] New comment on issue #5189: Remove \`set\_vertical\_scroll\` from displayio

Against the Adafruit bundle:

> grep -r set_vertical_scroll * | sort
libraries/drivers/displayio_sh1106/adafruit_displayio_sh1106.py:            set_vertical_scroll=0xD3,  # TBD -- not sure about this one!
libraries/drivers/displayio_sh1107/adafruit_displayio_sh1107.py:            set_vertical_scroll=0xD3,  # TBD -- not sure about this one!
libraries/drivers/displayio_ssd1305/adafruit_displayio_ssd1305.py:            set_vertical_scroll=0xD3,
libraries/drivers/displayio_ssd1306/a...
idle wharf
# 
cookiecutter cookiecutter-adafruit-circuitpython
Select target_bundle:
1 - Adafruit
2 - Community
Choose from 1, 2 [1]: 2
github_user []: myuser
author_name: Me Me Me
company []:
library_name: libraryname
library_description []:
library_keywords []:
library_prefix []: prefixname
adafruit_pid []:
requires_bus_device []:
requires_register []:
other_requirements []:
pypi_release [no]:
sphinx_docs [no]:
 patrickwalters@Mac-mini  ~/code/adafruit  ll Prefixname_CircuitPython_libraryname/*.py
-rw-r--r--  1 patrickwalters  staff  1150 Aug 20 13:45 Prefixname_CircuitPython_libraryname/prefixname_libraryname.py
jaunty juniper
#

it is a year old repository, could it be an old bug ?

idle wharf
#

Wait...

#

I said that wrong.

#

cookiecutter does drop the CP

#

OK. I get it. Sorry.

#

Since this is an existing year old library, I think we go with the PR and not make the library change.
Sound good?

#

I could make a PR to change the library and examples, but that would affect learn guides, and of their own docs, etc..

jaunty juniper
#

seems ok to me

idle wharf
#

done and done.

idle owl
idle wharf
#

Thank you @jaunty juniper for your help !

jaunty juniper
#

I find the rules of cookiecutter + circuitpython-build-tools to be strict and maybe hard to follow, but hopefully moving circup to the json file would give much more flexibility by letting the bundle maker specify any arbitrary pypi and repository name for a module

idle wharf
#

Yeah it should either be impossible to get it wrong, or very flexible.

idle owl
#

I did not.

#

Looking now.

idle wharf
#

Should "Community" now be CircuitPython_Org ?

idle owl
#

Not sure on that one, worth discussing.

#

I'm going to tag FoamyGuy on this issue, I think. Who else has been working on cookiecutter lately? Checking

#

FoamyGuy 😄

#

Tagged.

jaunty juniper
idle owl
#

We have a number of things to iron out in this transition, clearly.

idle wharf
#

I ran a test on all the libraries in circup show

Total Libraries Tested: 333

0 Failed libraries: []

4 Libraries with failed dependencies: [
{'adafruit_pyoa': 'adafruit_display_button'},
{'arrowline': 'vectorio'},
{'jepler_udecimal': 'astroid'},
{'pimoroni_circuitpython_ltr559': 'pimoroni_ltr559'} <-- this is now fixed
]

jaunty juniper
#

try the pyi name (which would be used in a requirements.txt file) circup install circuitpython-displayio-annotation

#

it doesn't work because it's not prefix_circuitpython_thing

#

so I don't know whatever rule cookiecutter has on that, but circup has not been updated to follow it ?

idle owl
#

Ok I'm supposed to have been out a few hours ago, which clearly did not happen. Have a lovely weekend all!

jaunty juniper
#

have a nice weekend 🙂

idle wharf
#

Thanks @idle owl This can wait... I just always pop-up about this time on Fridays.
have a great weekend!

idle owl
manic glacierBOT
tulip sleet
jaunty juniper
#

is there a version of vectorio usable on blinka ?

storm minnow
lone axle
manic glacierBOT
jaunty juniper
#

I thought adafruit_shapes would abstract that for blinka, but it doesn't

lone axle
manic glacierBOT
lone axle
#

Ah, I hadn't thought about using the shapes library for it. Perhaps there is some way we could make a minimal wrapper around display_shapes that would allow blinka code to use it for implementing the vectorio API.

lone axle
#

PyLint 2.10 was released a few hours ago, and it seems that it may have broken our actions run 🙁

ember iris
#

Ok, referring to the recent deep dive--this compression is the area of the core lib that I'm interested in but c is not what I'm good at. I have other projects, but would me focusing on attempting a compression implementation (within my means) be more valuable than me working towards supporting some other code, like the sgp40 voc algorithm? Both are things I'm really driven to work on, but their value and their probability of success is different. (Additionally I want to look at the qr codes and ‘flatten’ codes on a curved surface). Setting that aside, I don't want to take up time learning about the qstrings and compression from you if spending elsewhere would be more worthwhile. This is mostly a question because you're next to CP7.0's official release

tulip sleet
lone axle
#

indeed. I did verify that it's something related to that version. I installed it locally and my pre-commit fails the same way that actions did. So at least it's consistent. Trying to look into it further now to see if there is anything we can do to get it back to working.

#

If I understand the pre-commit config correctly it seems like we have pegged the pylint on library code to version 2.7.1 but the examples pylint does not specify a version so I assume it uses the latest.

slender iron
#

@ember iris I definitely want to talk with jeff about it next week

#

one of the tricky things with the compression is that its worth accounting for how much code space is occupied by the decompression code

lone axle
#

Okay I figured out how to fix it locally. There seem to be 2 things at play. This is the error:

PYLINTHOME is now '/home/runner/.cache/pylint' but obsolescent '/home/runner/.pylint.d' is found; you can safely remove the latter
Traceback (most recent call last):
  File "/opt/hostedtoolcache/Python/3.7.11/x64/bin/pylint", line 8, in <module>
    sys.exit(run_pylint())
  File "/opt/hostedtoolcache/Python/3.7.11/x64/lib/python3.7/site-packages/pylint/__init__.py", line 21, in run_pylint
    from pylint.lint import Run as PylintRun
  File "/opt/hostedtoolcache/Python/3.7.11/x64/lib/python3.7/site-packages/pylint/lint/__init__.py", line 76, in <module>
    from pylint.lint.parallel import check_parallel
  File "/opt/hostedtoolcache/Python/3.7.11/x64/lib/python3.7/site-packages/pylint/lint/parallel.py", line 8, in <module>
    from pylint.lint.utils import _patch_sys_path
  File "/opt/hostedtoolcache/Python/3.7.11/x64/lib/python3.7/site-packages/pylint/lint/utils.py", line 11, in <module>
    from pylint.config import PYLINT_HOME
  File "/opt/hostedtoolcache/Python/3.7.11/x64/lib/python3.7/site-packages/pylint/config/__init__.py", line 89, in <module>
    for filename in os.listdir(PYLINT_HOME):
FileNotFoundError: [Errno 2] No such file or directory: '/home/runner/.cache/pylint'

The first thing printed is indicating that the directory they use for caching has changed, and it's a warning that the old (now obsolte) directory still exists but can be deleted. Apparently it's existence is causing the new cache directory to not get created successfully which is what ends up making the final FileNotFoundError

ember iris
slender iron
#

don't let the C hold you up. you can get better with practice 🙂

ember iris
#

if I can get all three aspects below a target I'd be happy, if I can't prove I could in python, it's not worth while. That's the risk

lone axle
#

In my local environment I renamed (instead of deleting) the old now obsolete cache directory and then tried running pylint again and it succeeds.

slender iron
#

@lone axle no clue about the "Cannot read property..." error

lone axle
#

To fix it in actions I think we have 2 options 1) change the library code pylint hook to use the latest version 2.10.0 so that it will not create the old cache directory. Or 2) add rm -r /home/runner/.pylint.d to the actions before the examples pylint runs so that it will delete the old cache directory.

slender iron
#

can we bump the cache key?

tulip sleet
#

this is a library CI run

lone axle
#

I'm not sure what that means tbh.

slender iron
#

where is the old pylint cache directory coming from?

lone axle
#

I think it's getting created when the pylint (library code) hook runs because that one seems to be pegged to version 2.7.1

tulip sleet
lone axle
#

I could try either (or both) of the ideas that I think will fix it on the HID library. But if I understand correctly implementing either will ultimately require changes across all of the libraries.

tulip sleet
#

there's no global place where either version of pylint is pinned, is that right?

lone axle
#

Not that I know of. The two places that are relavent to where it comes from are 1) inside the .pre-commit-config.yml inside each library for the (library code) hook. And 2) inside of workflows build.yml inside each library where it runs pip install --force-reinstall pylint Sphinx sphinx-rtd-theme pre-commit which installs the latest release version and I assume is the one that is getting used by the (examples code) hook.

tulip sleet
#

I was going to ask you if you thought this was basically a bug in the new pylint.

#

So maybe they will fix this quickly.

lone axle
#

yep, I think the intended behavior might have been for it to warn you once (or perhaps on every run) about the old cache directory being no longer needed. But also intended for it not to actually fail to run if it does exist.

#

when I renamed the old cache dir it does run successfully though so I think the failure is somehow tied to that check that is trying to issue the warning.

tulip sleet
#

In the long run, tracking versions of things like pylint is a total pain. the .pre-config-commit.yaml requires specific versions, not latest, if I remember (I'm looking at the docs too)

manic glacierBOT
#
Docteh
[adafruit/circuitpython] New comment on issue #5192: Translations with two\-byte unicode characters shouldn't use two bytes for every character

I think the compression needs to see which method is more efficient. Most languages using code points with high values are probably using lots of code points with high values.

Throwing a few fancy apostrophes into the middle of many latin characters seems to have been unexpected.

The fancy apostrophe in question is 8217 in hex 0x2019
Also represented as three bytes in UTF-8

echo -n ’ |hexdump -C
00000000  e2 80 99                                          |...|
00000003
`...
tulip sleet
#

i'm trying to think of how to get around this. we could fork these tools, and set the tag to adafruit-choice, and keep moving it forward as needed. That doesn't really help the pip installs, but maybe we could do something there. We could alnteratley have the build.yml read the .pre-commit-config.yaml and pin its choices to those choices, but then we still need to advance the pre-commit choices at some point, and then we can't use the forked repo thing, because pip is looking on pypi

polar bobcat
#

Hello, I'm trying to code a hotkey for the macropad we got in the last adabox, I can't seem to program a button for the windows magnifier.
The shortcut in windows is 'Windows +'
I tried (0x004000, 'Magnify', [Keycode.WINDOWS, 'z']),

#

sorry I meant (0x004000, 'Magnify', [Keycode.WINDOWS, '+']),

manic glacierBOT
#
mocleiri
[adafruit/circuitpython] New comment on issue #4671: Native Tensorflow Lite support

I've built a micropython integration for tensorflow lite micro: https://github.com/mocleiri/tensorflow-micropython-examples

I've implemented all of the upstream examples except for magic wand. Some are just running micropython code with hard coded sample data and others like micro_speech can run in real-time with audio sampled over i2s.

I'm progressively expanding what has been done and I'm interested in integrating with circuit python to expand the number of people that can use it to ...

jaunty juniper
manic glacierBOT
manic glacierBOT
#
henoya
[adafruit/circuitpython] New comment on issue #5076: International keycode cannot be sent or received by macOS host

Dan-san

Thanks!
I have tried pico-kbd-0-255.uf2.zip and it works fine.
I have confirmed that keycodes that were not entered in the previous version are now entered!
The ￥ character is assigned to the same character code as / (backslash) in ASCII code, but it is not supposed to be displayed unless you have a Japanese display environment.
In my computer environment, INT1 (135) was correctly entered and the ￥ character was displayed.

I can now use the keyboard without resorti...

manic glacierBOT
manic glacierBOT
#
dhalbert
[adafruit/circuitpython] New comment on issue #5076: International keycode cannot be sent or received by macOS host

Great! Thank you for testing. This fix has already been merged as part of #5151, since I felt that the descriptor needed improvement in any case. It is already available in any "Absolute Newest" release, and will be in the next release, which will probalby be 7.0.0-beta.0.

The specific commit is https://github.com/adafruit/circuitpython/pull/5151/commits/38f42818e9a01226cc62d4d1ba7bc4dc264ff5b8.

#
jepler
[adafruit/circuitpython] New comment on issue #5192: Translations with two\-byte unicode characters shouldn't use two bytes for every character

The portion that you saw showing up as two bytes per character is the "dictionary", which is a small portion of the message data overall. (called "words" in genhdr/compression.generated.h)

It turns out this is not a savings for all languages, basically for the reasons @Docteh mentions above, we need make the compression system accommodate either situation and determine which is more efficient. (Another alternative would be to use an variable length encoding that indexes into the array ca...

lone axle
#

Is there a best practice for a "default value" for a uint32 argument to a function in core code? Or is there an existing usage somewhere I could reference? My first inclination was to use -1 but that didn't work since it's an unsigned integer. 0 is a valid value that the user could be legitimately passing so it can't be that either. My aim is to make the argument optional, so if the user passes it then the function will use what is passed, but if they don't then it will derive a value from some of the other non-optional arguments. Maybe it's possible to use None or null for this purpose somehow?

blissful pollen
#

I'll take a look in a few minutes but I think in one case I used MAX_INT or something that was unrealistic to ever occur

#

or I guess if there isn't a good "default" make it a required keyword?

analog bridge
lone axle
#

Thank you. I'll check this out

analog bridge
#

One more thing, if I understand this correctly... your function doesn't take negative integers so have you tried the following:

{ MP_QSTR_x2, MP_ARG_INT, {.u_int = -1} },
```and then check for negative values
lone axle
#

I attempted to change it over to int instead of uint32_t but that led to errors when I try to compare it to other values which are uint32_t. Perhaps it's possible to do that, but I was unable to figure out the right way.

manic glacierBOT
#
lesamouraipourpre
[adafruit/circuitpython] Issue opened: #5194 \`pixel\_luma\` parameter is not used in \`displayio\_colorconverter\_compute\_tricolor\`

The pixel_luma parameter of displayio_colorconverter_compute_tricolor in shared-module/displayio/ColorConverter.c is not used at all.

> grep -r displayio_colorconverter_compute_tricolor *
shared-module/displayio/Palette.c:        displayio_colorconverter_compute_tricolor(colorspace, pixel_hue, luma, color);
shared-module/displayio/ColorConverter.c:void displayio_colorconverter_compute_tricolor(const _displayio_colorspace_t *colorspace, uint8_t pixel_hue, uint8_t pixel_luma, ui...
tulip sleet
# lone axle That was my original idea, but the argument in shared bindings is a `uint32_t` s...

Here's an example from usb_hid.Device of taking an int arg, checking to see whether it's in a certain range (there are other validation routines that check min and max), and then assigning it to an unsigned value. You could also cast it explicitly from signed to unsigned:

    const mp_int_t usage_page_arg = args[ARG_usage_page].u_int;
    mp_arg_validate_int_range(usage_page_arg, 1, 255, MP_QSTR_usage_page);
    const uint8_t usage_page = usage_page_arg;
lone axle
#

Thank you

tulip sleet
# lone axle Thank you

Actually, there's a simpler way: mp_arg_validate_int_range et al return the mp_int_t, so you can do it in one go, e.g.

const size_t max_events = (size_t)mp_arg_validate_int_min(args[ARG_max_events].u_int, 1, MP_QSTR_max_events);
#

I added the return value later, I think.

empty gull
#

On the UM ESP32 S2 Feather when I go to sleep with an alarm can I have any effect on GPIO pins during sleep? For example, I would like the UART TX pin to, at a minimum, not float.

tulip sleet
#

The pins are disconnected to save power during deep sleep. You could add a mild external pull-up or pull-down to hold the level when it's unpowered

manic glacierBOT
#
FoamyGuy
[adafruit/circuitpython] Issue opened: #5195 Bitmap allows pixels to be set to color indexes that exceed value\_count

CircuitPython version

Adafruit CircuitPython 7.0.0-alpha.6 Adafruit PyPortal with samd51j20

Code/REPL

bitmap_test = displayio.Bitmap(2, 2, 3)
bitmap_test[0,0] = 3  # no error raised
bitmap_test[0,0] = 4  # raises ValueError "value requires too many bits"

# and another example: 

bitmap_test = displayio.Bitmap(2, 2, 5)
bitmap_test[0,0] = 15  # no error raised
bitmap_test[0,0] = 16  # raises ValueError "value requires too many bits"

Behavio...

manic glacierBOT
crimson ferry
#

how unusual would it be for gc.collect() to take something on the order of a minute to complete?

manic glacierBOT
lone axle
crimson ferry
#

thanks, I've never seen it take human-scale time either. I'm trying to isolate some hard faults and hangs on MagTag that seemed to increase when I put an I2C RTC on it, so I moved the RTC to the FunHouse and FunHouse got stuck somewhere (keyboard interrupt showed it at gc.collect() bute it should have been in a different place in the loop)

tulip sleet
#

it might be that tons of garbage is being created over and over for some reason, so it's spending most of its time collecting

#

i.e. it's multiple collects

#

maybe add a counter to the loop

crimson ferry
#

thanks, Dan, I'll keep that in mind as I watch this. normally, the loop works fine, flashing an LED and updating the clock on the display every second

tulip sleet
#

ah, so you can tell when the loop goes around, so if you have not already added print statements to see where it is, maybe do that. Or maybe I2C is getting stuck

crimson ferry
#

I'll add some debug prints, it's pretty intermittent. I do lean toward something I2C-related

tulip sleet
#

there are some very recent fixes to background processing; try an Absolute Newest build if you have not yet

crimson ferry
#

I do have a loop counter, and pretty much anything that happens gets printed with a timestamp, loop count, memory, and other goodies

manic glacierBOT
manic glacierBOT
#
dhalbert
[adafruit/circuitpython] Pull request opened: #5196 improve SAMD audio DMA

These are changes to audio DMA on SAMD to reduce scratchiness and other artifacts. The changes are similar to #5079: determining which buffer to fill is no longer done with a flip-flipping boolean, but instead by figuring out the buffer that finished playing in the DMA ISR. The code has also been restructured somewhat as with #5079. Also some extremely minor changes were made in a couple of RP2040 files.

On my PyPortal, I heard an uncommon trailing artifact when playing a A440 16-bit 32-kH...

manic glacierBOT
#
hierophect
[adafruit/circuitpython] New comment on issue #4671: Native Tensorflow Lite support

@mocleiri my first instinct would be the STM32H7 (probably H743), which we have basic support for, is used by OpenMV, and has a precedent for being used for this kind of application. However, a number of Circuitpython boards use external flash, including the ESP32S2, the i.MX RT, and the RP2040, all of which could theoretically support the firmware size you're mentioning. It'd probably be nice to pair it with a platform that supports a camera or audio features.

manic glacierBOT
manic glacierBOT
#
PeterFalken
[adafruit/circuitpython] Issue opened: #5198 Macropad RP2040 sometimes doesn't report position change\.

CircuitPython version

Adafruit CircuitPython 7.0.0-alpha.6 on 2021-08-12; Adafruit Macropad RP2040 with rp2040

Code/REPL

## Imports
from adafruit_macropad import MacroPad

## Global variables
os_index = 0
app_index = 0

l_position = 0  ## last position

## Init
macropad = MacroPad()
text_lines = macropad.display_text(title="MacroPad Info")

## Main loop
while True:
    diff = macropad.encoder - l_position
    l_position = macropad.encoder
 ...
#
dhalbert
[adafruit/circuitpython] New comment on issue #5198: Macropad RP2040 sometimes doesn't report position change\.

IIn the code above, trying getting the macropad.encoder value only once, since it may vary between the two calls at the beginning of the loop.

The rotary encoder handling is interrupt-driven, and sometimes can lose clicks when the encoder transitions happen too fast. I'm not sure that's the issue here. There is also a known problem about reversing: #3875. We've spent time tweaking the code, but there's a tradeoff between accuracy in terms of registering false transitions vs losing transi...

manic glacierBOT
empty gull
manic glacierBOT
#
amit-sides
[adafruit/circuitpython] Pull request opened: #5199 Adding quick refresh support

During the process of creating the ARAMCON 2 Badge, we needed the E Paper Display to refresh more frequently. Therefore, we added an API function that allowed us to change the start_sequence of an EPaperDisplay object and control it's waveform data that is used to update the display.
I'm sure this is not the prettiest way to do it, but we needed a fast solution so I decided to change the core base as little as possible.
I came out with the proposed solution that doesn't change any of the "d...

manic glacierBOT
#
PeterFalken
[adafruit/circuitpython] New comment on issue #5198: Macropad RP2040 sometimes doesn't report position change\.

After updating the code to only call macropad.encoder - the behavior is improved, less positions are lost if turning the encoder fast. I think that in the end the positional code should actually be something like this:

while True:
    n_position = macropad.encoder
    diff = n_position - l_position
    l_position = n_position
    if 0 != diff:
        if macropad.encoder_switch:
            os_index+=diff
            ## Reset app_selector
            app_index = 0
   ...
manic glacierBOT
manic glacierBOT
manic glacierBOT
#
PeterFalken
[adafruit/circuitpython] New comment on issue #5198: Macropad RP2040 sometimes doesn't report position change\.

Dan,
Let's close this issue, since the root cause is already being tracked on the ticket you mentioned on the first comment.
I'll submit a PR for the Macropad Hotkeys with some improvements. The feature addition - I'll work on it and get in touch once it's ready for release, I could probably take care of updating the guide as well.

Thanks for the help, I learned a couple of things.

manic glacierBOT
#
lesamouraipourpre
[adafruit/circuitpython] Pull request opened: #5200 Remove \`displayio\_copy\_coords\` from shared\-module displayio

The method displayio_copy_coords is identical to displayio_area_copy. Remove displayio_copy_coords as it is the lesser used and replace with displayio_area_copy

void displayio_copy_coords(const displayio_area_t *src, displayio_area_t *dest) {
    dest->x1 = src->x1;
    dest->y1 = src->y1;
    dest->x2 = src->x2;
    dest->y2 = src->y2;
}

void displayio_area_copy(const displayio_area_t *src, displayio_area_t *dst) {
    dst->x1 = src->x1;
    dst->y1 = src-...
manic glacierBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
#
mocleiri
[adafruit/circuitpython] New comment on issue #4671: Native Tensorflow Lite support

Thanks for your feedback.

I'm tracking circuitpython integration on: https://github.com/mocleiri/tensorflow-micropython-examples/issues/27

I'm looking at reducing the ops and firmware size in:
https://github.com/mocleiri/tensorflow-micropython-examples/issues/28

I'll start with the Cortex M7 boards like openmv and teensy 4.1 and then into the others you listed.

I'll ping back on this issue once there are firmwares to test.

#
FoamyGuy
[adafruit/circuitpython] New review comment on pull request #5145: Bitmaptools boundary\_fill

I am not 100% sure about the reason, it seems like displayio_bitmap_set_dirty_area might be inclusive on one end and exclusive on the other, I don't see anything documenting that though. And it's not consistent on every fill which definitely seems odd to me.

I did try initially without the +1s and some of the fills are not showing the far right and bottom edge lines of pixels being filled. Most of the logic was copied from the rotozoom function which does also have the +1s in it's ...

manic glacierBOT
solar whale
#

Anyone familiar with the new RAYTAC-nrf52840 dongle? https://www.adafruit.com/product/5199 With the pre-installed firmware, I can see it with the nrf-connect app and see changes in the button state and well as set the LED on/off. I tried uploading CP from tip of main to it and it does not seem to boot - I see several errors in dmesg related to the file system. Does this device need some special handling?

#

dmesg errors ```[439587.025403] sdd:
[439587.027901] sd 8:0:0:0: [sdd] Attached SCSI removable disk
[439607.466351] usb 3-3.2: USB disconnect, device number 8
[439607.481272] blk_update_request: I/O error, dev sdd, sector 11 op 0x1:(WRITE) flags 0x0 phys_seg 8 prio class 0
[439607.481307] Buffer I/O error on dev sdd, logical block 11, lost async page write
[439607.481311] Buffer I/O error on dev sdd, logical block 12, lost async page write
[439607.481313] Buffer I/O error on dev sdd, logical block 13, lost async page write
[439607.481315] Buffer I/O error on dev sdd, logical block 14, lost async page write
[439607.481317] Buffer I/O error on dev sdd, logical block 15, lost async page write
[439607.481327] Buffer I/O error on dev sdd, logical block 16, lost async page write
[439607.481329] Buffer I/O error on dev sdd, logical block 17, lost async page write
[439607.481330] Buffer I/O error on dev sdd, logical block 18, lost async page write
[439607.481344] blk_update_request: I/O error, dev sdd, sector 269 op 0x1:(WRITE) flags 0x0 phys_seg 8 prio class 0
[439607.481347] Buffer I/O error on dev sdd, logical block 269, lost async page write
[439607.481349] Buffer I/O error on dev sdd, logical block 270, lost async page write
[439607.481379] blk_update_request: I/O error, dev sdd, sector 517 op 0x1:(WRITE) flags 0x0 phys_seg 1 prio class 0

tulip sleet
solar whale
#

@tulip sleet Thanks -- I'l file an issue.

tulip sleet
solar whale
#

no drive -- I can connect to /dev/ttyACM0 and it reports connections. With nrf-connect app in the "client" tab I can set the LED on/off and in the "log" I can see the button status. I get a bootloader if I power it with button pressed -- I saved the current.uf2 and was able to restore it after trying CP.

manic glacierBOT
#
jerryneedell
[adafruit/circuitpython] Issue opened: #5201 Raytac nrf52840 Dongle not booting

I tried using one of the new RAYTAC-nrf52840 dongles? https://www.adafruit.com/product/5199 With the pre-installed firmware, I can see it with the nrf-connect app and see changes in the button state and well as set the LED on/off. I tried uploading CP from tip of main to it and it does not seem to boot - I see several errors in dmesg related to the file system. Does this device need some special handling?
dmesg errors

[439587.025403]  sdd:
[439587.027901] sd 8:0:0:0: [sdd] Attached ...
tulip sleet
#

ok, yes, I see the same thing. (I hadn't read the product description yet)

solar whale
#

I was just curious what I could do with it and had a few minutes to try it out. I have to go now, but will look at it again tomorrow,

#

if we still have power...

tulip sleet
solar whale
#

Cool! I'll try that tomorrow - maybe something recent broke.

manic glacierBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
#
dhalbert
[adafruit/circuitpython] Pull request opened: #5202 nrf: remove critical section around sd\_app\_evt\_wait\(\)
  • Fixes #5201.

#5182 added a critical section around sd_app_evt_wait(), which prevents USB from coming up (and perhaps other things).
https://github.com/adafruit/circuitpython/blob/fdfcaf0b72c61165922c63985909fef0b4868260/ports/nrf/supervisor/port.c#L367-L372
This removes that critical section, which was not present before.

Other uses I see of sd_app_evt() don't do anything special (no critical sections, etc.)

@tannewt I'm not sure if this the most correct fix, based on what yo...

manic glacierBOT
manic glacierBOT
#
jepler
[adafruit/circuitpython] New comment on issue #5192: Translations with two\-byte unicode characters shouldn't use two bytes for every character

I had an additional idea, which was to make the entries in "words" index "values" rather than being code points; the indexing is "utf8-like" if more than 0-255 indices in values are needed. This should be a win on average, as (no matter the language) most code points in "words" should be represented by just 1 byte. However, it ends up a bit worse for latin languages still.

language before after delta
fr 1328 1312 -16
de_DE 1052 1032 -20
j...
manic glacierBOT
manic glacierBOT
#
waiweng83
[adafruit/circuitpython] New comment on issue #5186: Print function doesn't print the GPIO state correctly\.

I think I've found out the root cause for this problem.
I've done some test using the code below:

import board
import digitalio
import time

led = digitalio.DigitalInOut(board.GP0)
led.direction = digitalio.Direction.OUTPUT

count = 0

while True:
    if count > 10000:
        led.value = 1
    print(count)
    count += 1

And what I got was the LED turns on way before the printed counter is > 10000.

Can I confirm that the print function is non-blocking a...

manic glacierBOT
manic glacierBOT
#
lesamouraipourpre
[adafruit/circuitpython] Pull request opened: #5203 Use the modified args to sort in displayio\.Group

Calling sort() on a Group hard crashed.
The pos_args were modified but the unmodified ones were passed onto mp_obj_list_sort()

This has been tested on a PyPortal with the following code:

import displayio
import board
import adafruit_display_text.label as label
from terminalio import FONT

group = displayio.Group()
for i in range(10):
    lbl = label.Label(FONT, text=f"Label {i}", x = 100 - (i * 5), y = 100 - (i * 5))
    group.append(lbl)
    
for layer in grou...
#
dhalbert
[adafruit/circuitpython] New comment on issue #5186: Print function doesn't print the GPIO state correctly\.

Can I confirm that the print function is non-blocking and calling the print function is actually just adding the value to a queue and it will only be printed after the previous value.

There's a buffer for USB output, and the output is also buffered on the host side. On the host side, the buffer can get very large, so you'll have some lag. When and if the host-side buffer fills up, the USB driver there will do a USB stall to prevent getting more input, and the CircuitPython side will not ...

manic glacierBOT
stuck elbow
#

ouch, how did I miss that

manic glacierBOT
idle owl
#

@tulip sleet I have a BLE CP question.

#

If you have a moment.

tulip sleet
#

sure

idle owl
#

Ok, so folks are apparently running into thinking they don't have to have any BLE code in code.py to have the board show up in the Bluefruit LE Connect app. So I'm updating the guide. Is the following the minimum you would need to have it show up? ```py
from adafruit_ble import BLERadio
from adafruit_ble.advertising.standard import ProvideServicesAdvertisement
from adafruit_ble.services.nordic import UARTService

ble = BLERadio()
uart = UARTService()
advertisement = ProvideServicesAdvertisement(uart)

while True:
ble.start_advertising(advertisement)```

#

That wouldn't do anything else, but would it at least show up in the app?

tulip sleet
#

it is only slightly longer and actually does something

idle owl
#

Hmm.

#

Shrug, ok. I'll include that instead.

tulip sleet
#

you could do

ble.start_advertising(advertisement)
while True:
    pass
#

that is fine for showing up

idle owl
#

Ok.

#

I'll include both I suppose.

tulip sleet
#

agree it's a smoke test

idle owl
#

Wait, your example says Used with ble_uart_echo_client.py

#

Is that required?

tulip sleet
#

no, the Bluefruit app has UART in it too. We could change the comment

idle owl
#

Because this isn't meant to go into being a full on example. Simply to show people what to include in their code so it shows up.

#

Ah.

#

Eh I'll include what I pasted to you instead.

#

With your suggestion

tulip sleet
#

I will submit a PR to the example saying you can use the app too.

idle owl
#

Ok thanks

manic glacierBOT
onyx hinge
#

@tulip sleet if you ever had a chance to look, was there anything interesting to us in that latest ARM toolchain update?

tulip sleet
manic glacierBOT
tulip sleet
#

@idle owl I think will take this away, it's old:

unknown.png
#

everything is always in development :), but I think now this feature is as stable as anything else

idle owl
#

@tulip sleet Where is that? In the page I linked?

tulip sleet
#

it was on the front page of that guide

idle owl
#

K on it

#

Oh, wait, you did it.

#

I did not read apparently.

#

Thank you 🙂

onyx hinge
#

I checked a random build (pygamer) and size was 8 bytes smaller with 10.3.

tulip sleet
idle owl
onyx hinge
#

SERCOM0_CORE huh I wonder why SERCOM0..7_CORE appear as strings in firmware.uf2 of pygamer

#

genhdr/qstrdefs.generated.h:QDEF(MP_QSTR_SERCOM0_CORE, 20, 12, "SERCOM0_CORE")
huh is a qstr

#

oh samd clock module

turbid radish
#

?serverinfo

digital shoreBOT
#
adafruit
Owner

adafruit#3230

Channel Categories

8

Text Channels

59

Voice Channels

6

Members

30575

Roles

35

ID: 327254708534116352 | Server Created ·
idle owl
#

What would I do with time.monotnic() to track time in minutes/hours from start of running code.

#

Trying to track how long this sensor runs before it fails. I already forgot when I started it, so good job me.

#

It's on a Pi, so it's not going to reload every time something writes to something on my Mac.

#

I mean minutes would be fine, I can do math later.

#

Nevermind. Used time.time(), strftime and gmtime.

#

Seems to be working. Doubt it translates to pure CircuitPython though. But I don't need it to.

slender iron
manic glacierBOT
onyx hinge
#

<@&356864093652516868> we'll meet in about an hour -- please add your reports to https://docs.google.com/document/d/10fzYN9LFHmhD-v47O-HZw6WN8us9PmUtQTcjUbbHxEU/edit?usp=sharing and note if I'll need to read them

Google Docs
CircuitPython Weekly for August 23, 2021

CircuitPython Weekly for Welcome to the CircuitPython Weekly meeting notes! Feel free to add your Hug Reports and Status Updates early. During the meeting, we go through them as a round robin sorted by username. If you can’t make the meeting and would still like to participate, add your notes...

analog bridge
#

pca10100 de_DE fr builds are running out of space.

manic glacierBOT
onyx hinge
analog bridge
#

does it have CIRCUITPY_FULL_BUILD enabled?

analog bridge
#

just checked CIRCUITPY_FULL_BUILD is enabled on pca10100... should it be turned off or should I turn off atexit?

manic glacierBOT
onyx hinge
#

@analog bridge either way is fine

analog bridge
tulip sleet
#

it says these are on: bleio, analogio, atexit, audiobusio, audiocore, audiomixer, audiopwmio, board, busio, digitalio, errno, getpass, math, microcontroller, onewireio, os, pulseio, pwmio, rainbowio, random, rotaryio, rtc, storage, struct, supervisor, synthio, time, touchio, traceback, usb_cdc, usb_hid, watchdog

#

synthio could go too

analog bridge
#

building with synthio turned off now.

tulip sleet
#

@slender iron are you changing the apostrophe in that French translation in weblate? I could try but I have never used its web editor before.

slender iron
#

I haven't looked at it. I think jeff was going to

tulip sleet
#

@onyx hinge do you have time to look at that/

onyx hinge
#

It looks like that was not changed back by weblate

#

@analog bridge seemed to be indicating the problem was the new atexit module, rather.

tulip sleet
#

well, that's good, I only see the pca10100 failing atm

#

thanks for looking

onyx hinge
#

@slender iron did you find out whether you would be around to host the meeting on september 7 (labor day week) or whether we should cancel one?

solar whale
#

Jeff sounds like Scott

idle owl
#

We'll just have to deal with noise. 🙄

analog bridge
#

turning synthio off worked 🙂

lone axle
manic glacierBOT
slender iron
#

nothing wrong with a long list! kattni does a ton

uncut nexus
#

I do photoshop regularly

blissful pollen
#

ive done some as well, depending what needs to be done with it

uncut nexus
#

You can send me the image and what you need done.

modern wing
#

I used to be very much into photoshop (pre CS days), but it's been a while since I've tackled a project with it.

idle owl
#

Ok, so for the taco, it's a white outline on a black background. And it's kind of anti-aliased, which is to say the edges are a little fuzzy. It needs to be cleaner than it is, as in the edges need to be clearer, not as "soft".

pearl notch
#

What is the CircuitPython Code Editor interface?

idle owl
slender iron
#

melissa is actively working to make it much better

idle owl
#

@uncut nexus I DM'd you the 🌮

pearl notch
#

Looks interesting, @gilded cradle!

gilded cradle
tidal kiln
#

a walking palindrome 🙂

uncut nexus
#

@idle owl - yep got it thanks!

ember iris
#

Threads in the livebroadcast chat might be tough for the broadcaster to keep track of the conversation, as they probably aren't free to click around and catch up on 'more' chats

solar whale
#

Discord may not listen , but likely DHS does 😉

# 
while True:
    jpeg = cam.capture(jpeg_buffer)
    print(f"Captured {len(jpeg)} bytes of jpeg data")

    # b2a_base64() appends a trailing newline, which IO does not like
    encoded_data = binascii.b2a_base64(jpeg).strip()
    print(f"Expanded to {len(encoded_data)} for IO upload")
    io.publish(feed_name, encoded_data)
    print("Waiting 10s")
ember iris
#

Is that reboot something that takes place at a consistent time?--like after the same length of time

modern wing
#

Thanks everyone!

pearl notch
#

Good to be with you!

slender iron
#

Thanks all. Thanks for hosting Jeff

modern wing
#

You're doing just fine 🙂

pearl notch
#

Great job, @onyx hinge

modern wing
#

A machine that makes tacos out of thin air?

ember iris
# ember iris

Average time between reboots: 7458 seconds, but there's some error so give or take 50 seconds

#

Almost exactly once every two hours

modern wing
#

Hmm. Sourdough.

solar whale
#

👋

manic glacierBOT
ember iris
#

I'm going to see if I get a reboot with the co2 sensor disconnected. It'll still send dummy packets to my server using the wifi module. That should at least help reduce the problem space

onyx hinge
#

@solar whale it's shifted 1 byte ```>>> codecs.decode(b"///Y/+AAEEpGSUYAAA==", "base64")
b'\xff\xff\xd8\xff\xe0\x00\x10JFIF\x00\x00'

codecs.decode(b"/9j/4AAQSkZJRgABAA==", "base64")
b'\xff\xd8\xff\xe0\x00\x10JFIF\x00\x01\x00'

solar whale
#

Ah -- Thank you --that is a huge discovery....

#

If all else fails -- I can add the 0xff back in by hand!

idle owl
#

@lone axle So two things. Adding the other bundle(s) to the Project Bundler isn't going to be much work for Justin, as long as they have the same layouts and include the .json file. Also, I'm going to email intro you to Justin so you can discuss the code - it's apparently well tied into Learn and is gnarly due to 6x/7x support being needed, so extracting it out for the purposes of sharing it between things is not going to happy anytime soon. That said, Justin is happy to discuss it with you so if you'd like to work on something to share between the screenshot maker and Circup, you can at least know what to aim for so that once it's feasible to open up the bundler (once 7x is stable), sharing between all three will be a much easier process. Sound good?

lone axle
#

Yep, sounds good to me. Thanks again.

idle owl
#

@lone axle Are the other bundles the same, in that, do they have the same features?

manic glacierBOT
lone axle
# idle owl <@!382939733107408897> Are the other bundles the same, in that, do they have the...

I believe they are. When I started the CircuitPython org bundle I started essentially from a copy of the existing bundle and only changed what was necessary to have different submodules inside of it.

If there are outstanding differences I'm willing to work on normalizing them because I definitely do think it will be easier for all of the tools to work with each of them if they do as much as possible the same.

idle owl
jaunty juniper
#

there is a difference in the convention used to name the repositories/pypi names, but that is not an issue using the json file

#

the circuitpython org bundle doesn't seem to follow the prefix-circuitpython-library format

lone axle
lone axle
idle owl
#

Hmm. I'm not sure what the Project Bundler does, so I don't know whether this will be an issue or not.

onyx hinge
#

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/1bO1IU2lth51Bl0pKTlVy9IYak9HuB7GFfyaRQhFISxs/edit?usp=sharing

Google Docs
CircuitPython Weekly for August 30, 2021

CircuitPython Weekly for August 30, 2021 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...

onyx hinge
lone axle
#

We can change anything needed if it pops up as an issue. There aren't many in the org bundle yet, so it would not be such a big hassle to change them now if needed.

idle owl
#

Yep. Sounds good.

ember iris
onyx hinge
#

@ember iris yeah that's the one.

#

I looked at it for a bit but wasn't able to tease out a cause or fix, didn't get much of anywhere unfortunately.

idle owl
#

MP3 playback worked on the Pico. This little speaker is loud.

ember iris
#

This is way outside my area of experience, but
When the esp makes a 'connect' call to the router, does is do anything related to the time of the call? Like does it compare it's clock to the routers clock?

manic glacierBOT
idle owl
#

Running on borrowed time now. The forum post said it failed after "several hours". ```Elapsed: 03:17:19

Data Available!
CO2: 1974 PPM
Temperature: 24.12 degrees C
Humidity: 42.31 % rH

Waiting for new data...```

manic glacierBOT
idle owl
#

Also, those levels. Oof.

ember iris
idle owl
#

It doesn't say anything about trying a different power supply.

#

So still 5 more hours to be on borrowed time for this issue.

ember iris
# idle owl Also, those levels. Oof.

I'll see if my setup fails with no co2 sensor plugged it. If there's no sensor and it doesn't fail, maybe it's related to the co2 sensor. If there's no sensor and it does fail maybe my problem is related to the wireless

idle owl
#

@ember iris This isn't the same sensor. This is the SCD-30.

manic glacierBOT
idle owl
#

But it's still worth testing for your situation.

#

(This is also on Pi.)

ember iris
#

which is reliable

idle owl
#

I'm using the "official RPi power supply" only because I couldn't find anything else not already in use that was 3A.

#

Which the Pi 4 apparently needs.

slender iron
idle owl
prime flower
ember iris
ember iris
solar whale
manic glacierBOT
ember iris
#

Additionally, because this issue is pretty predictably once every two hours, I predicted when I'd see it next, and I didn't get the edited code in time. I was 3 minutes too late. So now I have to wait another 2 hours

manic glacierBOT
prime flower
#

I have some ideas but need to see code

tulip sleet
#

@onyx hinge don't merge 5145 yet, I am waiting to make sure #5206 is ok, and then I will merge against that

#

@slender iron &^^ (sorry I misread and thought Jeff merged it)

manic glacierBOT
idle owl
#

@tulip sleet Question. This code. (I can't test it right now because my partner is in a meeting and it is LOUD.) Does this keep repeating the two files over and over, or does it play them each once and stop? ```py
mp3files = ["slow.mp3", "happy.mp3"]

mp3 = open(mp3files[0], "rb")
decoder = audiomp3.MP3Decoder(mp3)

while True:
for filename in mp3files:
decoder.file = open(filename, "rb")
audio.play(decoder)
print("Playing:", filename)
while audio.playing:
pass```

unique roost
#

hello friends, how you all doing? I have a question: I've been trying to use the aesio module on my ItsyBitsy M4 but so far no luck (it fails right at the python import statement). If aesio is supported on this board, how do I get it to work? Much appreciated.

idle owl
#

I think it loops. If I took it out of the while True: does it play them once each and stop?

tulip sleet
#

right, because the for is inside the while True

idle owl
tulip sleet
#

yup

idle owl
#

Thanks!

tulip sleet
#

Perhaps we should include it on more boards. What is your use case?

unique roost
#

danh, thank you for you reply. I would like to experiment creating an encrypted file to hold passwords, such that I could send a command from my console (serial) and have the microcontroler "type" the chose option into the host (via hid). Again, it is just an experiment for a proprietary password holder. 🙂

#

@tulip sleet One more thing: if that could be made available, what I was planning on doing would work even better on the Titano because of that nice display, where the commands could be selected and sent to the python code.

idle owl
#

@onyx hinge Are you still around? If not, my question can wait until tomorrow.

onyx hinge
#

@idle owl go ahead

idle owl
#

@onyx hinge How would you describe this? ```py
decoder = audiomp3.MP3Decoder(open("slow.mp3", "rb"))

audio.play(decoder)
while audio.playing:
pass```So far I have "Next, you create the decoder object and tell it the name of the MP3 file you'd like to play, in this case, "slow.mp3"." But I don't know how to describe the audio.play() line.

onyx hinge
#

the audio.play line starts actually playing the decoded mp3 on the speaker or headphone

#

how do you say it about wave files?

idle owl
#

Ok, that's what I was feeling, but that felt too simple.

#

Um... not sure.

#

That works though.

#

Thanks, Jeff!

idle owl
#

@onyx hinge Ok, one more if you're still around. Otherwise, it, too, can wait until tomorrow.

onyx hinge
#

still conscious

idle owl
#

Valid state.

#

Ok, so this code py mp3 = open(mp3files[0], "rb") decoder = audiomp3.MP3Decoder(mp3) Are you opening the first MP3 file in the list for the purposes of creating the decoder object, which is then used later with decoder.file = open(filename, "rb") to run through the list of filenames?

#

Or am I not following what's going on there.

#

Full snippet: ```py
mp3files = ["slow.mp3", "happy.mp3"]

mp3 = open(mp3files[0], "rb")
decoder = audiomp3.MP3Decoder(mp3)

for filename in mp3files:
decoder.file = open(filename, "rb")
audio.play(decoder)
print("Playing:", filename)
while audio.playing:
pass```

onyx hinge
#

yeah, it's a bit of a defect in the API

#

if you are getting fancy to re-use the decoder object, you still have to construct it with some mp3 file, it doesn't matter which one

#

then each time before you play you set its file property to the one you actually want to play next

idle owl
#

Ok

#

Thanks

ember iris
idle owl
ember iris
#

I'm going to keep testing as well. Next test is just to plug it back in and see if it happens, followed by swapping the stemma qt plugin I was using, followed by adjusting the time between i2c sensor reads. If I spot anything that might be a 'larger than me' issue I'll keep you posted, but it's probably just a me issue for the time being

idle owl
#

I can set it up with an ESP at some point though once I'm done doing my initial testing for the guide.

ember iris
#

If I can't narrow it down and if it keeps being buggy when I try a different board that'd be a huge help, but testing on a non-esp is probably best to make sure we're more easily able to pin down exactly what the problem is

idle owl
#

FunHouse is ESP with STEMMA, right? I mean I have a million ways to connect things, but I'm lazy 😄

ember iris
#

haha I've only got a few, but I've got a soldering iron so I can just make more!

idle owl
#

I even have some of those STEMMA to grabby-hands connectors. Used them the other day, so nice.

#

They went in and out of stock pretty quick I think. Easy enough to make them though if you have the grabby-hands.

#

Anyway, regardless, I'll give it a test. Up to 6.5 hours/8 on the SCD-30.

#

So glad I thought to put time code in it. Otherwise it would probably fail over night and I'd have no idea how long it ran.

#

@tulip sleet Disconnected you from the voice chat.

ember iris
#

Relatedly to the CO2 sensors, are you keeping yours within 10 feet of you and at desk level? I noticed if I was next to mine it would read 'dangerously' high, but if I was in the other room it would settle to around 400-700ppm

idle owl
#

I would have to setup data logging to see what happens when I'm not around.

#

Definitely worth checking, though, thank you.

ember iris
#

I'd be willing to be it's way lower over the full room, you're just near it and that's making a locally high reading

idle owl
#

I sure hope so, because, oi.

#

It drops some when the fan is blowing over me/my desk, but still not that much.

#

Once I'm done testing this for the clock-stretching failure, I'll set it up to data log.

#

Or I could set up like... the FunHouse with the display, to show the low and the high value.

#

At least I assume CircuitPython could do that with the appropriate code.

#

Or maybe I finally use my AIO account for something 😄

ember iris
#

Soo many options for fun data 😄

idle owl
#

I ended up with a ton of data for it though. Including all the buggy data where my freezer was hotter than the sun because of an unsigned int or something 😄

#

This was after Limor told me to put it in the freezer, but before she subsequently told me not to put it in the freezer because micros don't like freezing temps. The micro and the sensor were fine, it was the code that was unhappy.

ember iris
#

Unsigned int in a data log? Yeah that's a true logging project then! That, malformed strings that mess with the parser, etc

idle owl
#

Nice 😄

ember iris
#

Huh. I haven't thought about testing how the code runs on these at cold temps...

idle owl
#

I think for the most part, it's fine, but if you're powering it, maybe unplug it while it warms up when you finally remove it from cold temps. Unless it's gradual, in which case it's fine, but you can end up with condensation. Which is bad. But if left unplugged for a bit, it's fine.

#

Really, it's probably fine either way, but I think I was pushing the limits of the whole thing putting a little lipo-Feather-sammich in the freezer.

#

Nothing bad happened, but.... heh.

#

OK, I need to head out for dinner. Have a good one!

manic glacierBOT
#
dhalbert
[adafruit/circuitpython] New tag created: 7\.0\.0\-beta\.0
orchid basinBOT
tulip sleet
#

@slender iron @onyx hinge there is one board whose build is broken due to size, but it's an older version of a low-download-count board. I will fix that post-release.

slender iron
#

kk

onyx hinge
#

thanks dan!

manic glacierBOT
#
jepler
[adafruit/circuitpython] New comment on issue #5192: Translations with two\-byte unicode characters shouldn't use two bytes for every character

I doodled around with ideas to reduce the size of the huffman table in the ja translation, but I don't think there ends up being much gained, or maybe even a net loss.

General idea: Split code points into 8:8 values. Find out the number of distinct upper octets (70 in the case of the ja translation today), and store items as a combination of the lower 8 bits plus an index for the upper bits. In this case, though, only 1 bit per huffman table is saved encoding in this way, but you have to p...

spare prawn
#

I am getting a "TypeError: function does not take keyword arguments" when using UART.readinto whem specifying nbytes. I then tried passing the parameter nbytes not as a keyword and got the following error: "ValueError: length argument not allowed for this type". I took a look at /circuitpython/shared-bindings/busio/UART.c. I am not sure I understand the code but its not clear whether this is implemented. I am trying to build a generalized set of stream handling functions and found that I had to modify nordic.py to get read_into working. In this case its not clear to me how to work around using python and without recompiling. Any suggestions?

jaunty juniper
onyx hinge
#

you can use memoryview(buf)[:nbytes] or similar, to pass a view of your target buffer of the desired size

manic glacierBOT
spare prawn
#

@onyx hinge ok, thanks, i will try that. @jaunty juniper i just noticed that the first hit on google to read the docs was pointing me to v1...now see that note. thx!

orchid basinBOT
manic glacierBOT
lone axle
#

If I understand correctly this description for the pin parameter is no longer correct. This takes a pin object like board.D13 so it's not a PWMOut object is it?

unknown.png
#

And the sentence before the parameters could be updated to something like "Create a PulseOut object associated with the given pin."

tulip sleet
manic glacierBOT
tulip sleet
#

If you are worried about someone looking at CIRCUITPY, you can hide it by turning off the CIRCUITPY drive.

manic glacierBOT
manic glacierBOT
#
jpconstantineau
[adafruit/circuitpython] Issue opened: #5211 Rotaryio: enable user to initialize more than 4 encoders

I have a small macropad that has 5 rotary encoders. It appears that CircuitPython has a hardcoded limit of 4 rotary encoders to be setup on the nRF52840. (I use a BlueMicro840).

I created a test program to track the position of all 5 encoders. In the first iteration of the program, I initialize 5 encoders in sequence 1,2,3,4,5 and the 5th one doesn't respond.
In the second iteration, I moved the initializer for encoder ...

manic glacierBOT
manic glacierBOT
manic glacierBOT
#
microDev1
[adafruit/circuitpython] New comment on pull request #5208: Small memory savings

Looks like they do tend to use just %p. 0x is added afterwards if not already present.

case 'p':
    sprintf(buffer, "%p", va_arg(vargs, void*));
    assert(strlen(buffer) < sizeof(buffer));
    /* %p is ill-defined:  ensure leading 0x. */
    if (buffer[1] == 'X')
        buffer[1] = 'x';
    else if (buffer[1] != 'x') {
        memmove(buffer+2, buffer, strlen(buffer)+1);
        buffer[0] = '0';
        buffer[1] = 'x';
    }
    WRITE_BYTES(buffer);
    break;

...

manic glacierBOT
manic glacierBOT
#
jpconstantineau
[adafruit/circuitpython] New comment on issue #5211: nRF52 RotaryIO: raise error when more than 4 encoders are used

That makes sense and explains why performance is much better than in the arduino (Adafruit_nRF52_Arduino) software encoders.
Is there any other functionality that would consume channels as well? i.e. Does the consumption of GPIOTE channels impact other functionality? Take for example alarm.pin? Does that use the GPIOTE or pin SENSE mechanism?

Would there be any chance to add the hardware Quadrature decoder (QDEC) on the nrf52 for the first encoder and switch over to GPIOTE to enable 5 e...

lone axle
#

@idle owl After running through that list of changes I did find a few Learn Guide projects that are currently importing _pixelbuf and also a few that have created a PulseOut using the old API that takes a PWMOut object instead of the new one that takes the pin directly. I will be occupied with other things today, but I'll get a PR worked up tomorrow to resolve those.

manic glacierBOT
manic glacierBOT
#
urish
[adafruit/circuitpython] Issue opened: #5212 USB output only starts after main\.py finishes \(RP2040\)

CircuitPython version

- Adafruit CircuitPython 7.0.0-beta.0 on 2021-08-24; Raspberry Pi Pico with rp2040
- Adafruit CircuitPython 6.3.0 on 2021-06-01; Raspberry Pi Pico with rp2040

(tested on both versions)

Code/REPL

print("Hello, Pi Pico!")

Behavior

I'm working on adding CircuitPython support to my Raspberry Pi Pico simulator on Wokwi.com.

The USB Serial output seems to initialize only after main.py finishes running, so the user can't se...

manic glacierBOT
#
jepler
[adafruit/circuitpython] New comment on issue #5211: nRF52 RotaryIO: raise error when more than 4 encoders are used

I think I spoke dismissively about the above-linked algorithm once, but I took the time to actually test it and I think it may be a correct algorithm for quadrature decoding. I believed that any algorithm triggered only by edges from one of A or B had to be wrong, because this particular realization of an A-triggered quadrature decoder is wrong.

There are several differences between the proposed algorithm and mine, including that it i...

idle owl
lone axle
manic glacierBOT
#
dhalbert
[adafruit/circuitpython] New comment on issue #5212: USB output only starts after main\.py finishes \(RP2040\)

The behavior you describe does not happen on a real Pi Pico when running CircuitPython. USB is started when code.py starts.

Do you by chance have both a main.py and a code.py? Both are accepted: code.py is preferred for CircuitPython. (The accepted files are code.txt, code.py, main.txt, main.py, searched for in that order.)

Could you describe how you emulate the Pi Pico? Can you trace the USB activity?

idle owl
ember iris
#

@idle owl have you run the scd41 for more than 2 hours with the example code--the loop that check if scd4x.data_ready?

manic glacierBOT
#
urish
[adafruit/circuitpython] New comment on issue #5212: USB output only starts after main\.py finishes \(RP2040\)

Thanks for the speedy response and for trying to reproduce this on the real Pi Pico!

In this case, I only have main.py (also tried now with code, getting same behavior):

>>> import os
>>> os.listdir('/')
['main.py', 'diagram.json', 'boot_out.txt']

The Pi Pico is emulated using an open source library that I created for the purpose. I can trace USB activity, as well as connect with GDB to debug the CP code that is running inside the simu...

#
jpconstantineau
[adafruit/circuitpython] New comment on issue #5211: nRF52 RotaryIO: raise error when more than 4 encoders are used

Counting twice as fast may be an issue, especially with encoders that have mechanical detents (clicks). In such case, we would have double-counts per detent. We would need some kind of divider.

@jepler is your new algorithm the same as the one implemented in the arduino nrf52 stack? This is the one I was referring to in my comment above that didn't perform as well as what's curre...

#
dhalbert
[adafruit/circuitpython] New comment on issue #5212: USB output only starts after main\.py finishes \(RP2040\)

So if main.py is an infinite loop, you never see the EP0_IN_BUFFER_CONTROL write?

Both MicroPython and CircuitPython use TinyUSB as the USB stack. CPy calls tud_task() to handle pending USB work in the background. We check for background tasks at safe points in the VM interpreter via RUN_BACKGROUND_TASKS, aka supervisor_run_background_tasks_if_tick(). We see whether at least one 1/1024sec tick has elapsed.

You would probably want to watch for the execution of various routines in `...

idle owl
ember iris
idle owl
ember iris
idle owl
#

@onyx hinge This MP3 code isn't doing what I think it should be.

#

wait.

#

Let me try one thing.

#

Nevermind.

#

Forgot the MP3 fix wasn't in the last release, and had only tested 1 MP3 until this moment on Pico.

#

Updated CP and fixed it. Plays two now and doesn't hang.

ember iris
#

Was there an bug with gc recently?

tulip sleet
#

that was fixed? not that I remember

ember iris
#

Ok, I'll keep poking, then update to the beta release since I'm a few days behind. Just curious if I missed an obvious cause of the problem

#

Thank you!

idle owl
#

Ugh bummer, no strftime in CircuitPython, then?

jaunty juniper
#

you would use adafruit_datetime I believe ?

idle owl
#

I spoiled myself with using it on RPi.

#

Oh!

#

Thanks!

#

Hmm. We never setup docs for datetime.

manic glacierBOT
#
fasteddy516
[adafruit/circuitpython] Issue opened: #5213 No longer able to create custom USB HID devices\.

CircuitPython version

Adafruit CircuitPython 7.0.0-beta.0 on 2021-08-24; Adafruit Metro M4 Express with samd51j19
Adafruit CircuitPython 7.0.0-beta.0 on 2021-08-24; Adafruit QT Py M0 with samd21e18

Code/REPL

import usb_hid

# fmt: off
# This is only one example of a gamepad descriptor, and may not suit your needs.
GAMEPAD_REPORT_DESCRIPTOR = bytes((
    0x05, 0x01,  # Usage Page (Generic Desktop Ctrls)
    0x09, 0x05,  # Usage (Game Pad)
    0xA1, 0...
idle owl
#

@ember iris Here we go! ```Elapsed: 00:00:34

CO2: 818 ppm
Temperature: 23.2 *C
Humidity: 45.0 %

#

We'll see if the time math is valid, heh.

ember iris
#

Awesome! I think I might actually be hitting a memory error issue, but I'm waiting to see. I've just passed the 2 hour mark and it didn't crash, so now I'm waiting for the free memory to hit 0 to see if it crashes then

idle owl
#

I should actually set this up to data log and plug it into a power source instead of my laptop because it'll reboot randomly when things write, I think.

ember iris
#

Additionally, if nothing else I'll be able to fork the library and add the single shot sample for the scd41, (but I'm going to have to ask that you test it since I only have the 40) so while I went down a rabbit hole, hopefully I'll be able to have something to show for it 😅

idle owl
ember iris
#

Well together we have both, so we should be able to figure that out one way or another. I'm hoping it'll throw and error on mine and that'll be how we tell

idle owl
#

Good plan.

#

@ember iris Yeah it rebooted. Bleh.

#

Ok, setting up to be standalone.

ember iris
#

Oof. Well I'm increasingly confident the issue isn't with the sensor, it was just the sensor was what I was playing with when it happened. So on the plus side we might be able to eliminate that from a concern. I have another hour and 40 minutes before the gc needs to be called, and after that I can write a program to just eat up memory. If I get a reset then, we can be sure that is the cause, then I can upgrade to 7 beta, and try it there and see if it's gone. I just haven't upgrade because I wanted to know 'what' was causing my issue

idle owl
#

This rebooted because MacOS likes to reboot boards.

#

Not because of the sensor. IMO.

ember iris
#

I'll be honest, I thought the whole mac rebooted, so I didn't think it was the sensor just because I misread it 😅

idle owl
#

Hah!

#

Fair enough

ember iris
#

Is there a way to get issues on all sensors from a specific family, in this case sensirion? (Mild thought with no urgency--I'm still working on the sgp40 voc algorithm and the scd41 single shot made me notice their datasheets are consistent so I might be able to start queuing up some of those issues and seeing if I can't try some of them)

idle owl
#

The SCD-30 reset. Crap. It did not error though.

#

Apparently.

#

Says 7 minutes elapsed now though. Buffer is too short to see where it reset to 0.

ember iris
#

Hmmm

idle owl
#

Ran to 24 hours at least.

#

Maybe it resets to 0 at 24!

#

that would make sense with the way the code is probably written.

ember iris
idle owl
ember iris
idle owl
#

I can't view the code without stopping it, but I bet that's the issue.

#

I don't think it actually reset.

#

Because it's still running.

#

But the counter was hours, minutes, seconds. I didn't think about the fact that I should have added days. 😄

ember iris
#

Do you have a try/except for the I/O error?

idle owl
#

Nope. I wanted it to fail so I had eyes on it actually failing.

#

This is the basic simpletest code with some time code added to it.

ember iris
#

Sweet! I think it was just a rollover of the counter as well

#

So the scd4x said it needed a low noise power supply, does the scd30 have the same requirements?

idle owl
#

No idea!

ember iris
idle owl
#

OK, datalogger running, powered from the wall now.

#

If I was at all savvy, I'd be having it logging to AIO so I could track it remotely.

#

Alas...

#

Wait.

#

Would need a wifi board anyway. This is not one.

#

So my trusty little red LED will have to do the work for me.

ember iris
idle owl
#

Hmm.

#

The clock stretching thing is a legit concern though, as far as I knew.

#

But it's always clocks, so who knows.

#

Either way, I have to reply to that forum post saying that it's been running for over 24 hours successfully.

#

Bleh.

#

Was genuinely hoping for it to fail.

ember iris
#

I was looking for what threw the error and it's not in the scd30's source code. What are the odds their wiring is really long on top of clock stretching? Is it possible to ask for a picture of the wiring setup?

idle owl
#

Of course. We do it all the time.

ember iris
#

Yeah I've been hoping for my sensor to fail as well. If you can cause a failure, you can fine a solution that should work for them as well

idle owl
#

There was no reply to my reply, so I'm imagining this person figured out a workaround or the answer from the other person was sufficient.

#

I'll reply again, and if you want to reply and ask for a picture of the wiring, that's fine.

ember iris
#

Will do, if nothing else it'll be practice for me getting more involved which is a net win to me

manic glacierBOT
idle owl
#

BBIAB. Food.

ember iris
idle owl
manic glacierBOT
ember iris
#

Sensor disconnected way earlier than I expected !! D:

But I do think it's tied to memory, so I can just start eating memory to validate that

idle owl
#

@ember iris The person on that PR replied about the delay not being necessary. Should I reply and ask them to remove it then?

#

I find unnecessary delays come back to bite you later when troubleshooting something else and wondering about the delay being there. But if there's a delay on the other one, I guess maybe it's fine to merge it as is?

manic glacierBOT
ember iris
#

I was actually about to ask if you should, here's the current reply draft

manic glacierBOT
#
jepler
[adafruit/circuitpython] New comment on issue #5092: PWM\-generated tones may be off\-frequency at higher frequencies 
    uint32_t best_numerator = 0;
    uint32_t best_denominator = 0;
    uint32_t best_error = system_clock;
        
    for (uint32_t denominator = 0xffff; denominator > 0; denominator--) {
        uint32_t numerator = (denominator * sample_rate) / system_clock;
        uint32_t remainder = (denominator * sample_rate) % system_clock;

For sample rates above around 65536, the math overflows at (denominator * sample_rate)

ember iris
#

Personally I'd err on the side of following the datasheet, and not having a delay. But that's just because I usually use datasheets as my mental 'square one' so if I have a bug I can't easily solve, I go back to that to make sure I can follow along with the code.

But you're right, a delay doesn't hurt unless you're doing really time sensitive stuff which I doubt a co2 sensor would be involved in.

Still the sensor reads have a built in delay with if scd4x.data_ready, so I think it would be safe to remove this delay.

As to whether or not to keep the delay in, or remove it, I'm going to have to deffer to someone else here.

It's rambly so I'm editing it, but I'd remove the delay if there's no reason to keep in personally

manic glacierBOT
ember iris
#

I just am getting use to the style of circuit python, what is prioritized over what, so I'm not going to make a call with confidence yet

idle owl
#

OK, that sounds good. Clean up the ramble and reply. I wouldn't hedge your bets there though. I think it's fine to definitively request they remove the delay. I can reply after you do and agree.

idle owl
ember iris
idle owl
#

So since I have no way to verify what's going on with my sensor, I think I'll set an alarm for 5 hours from when I started running it, and take a look at the data at that point to see if it reset at any point.

#

@ember iris Ok requested changes on the PR.

#

@ember iris Also, if you haven't already, please comment on the feature request issue that you're looking into it. Good habit to get into.

#

You did!

#

Nevermind

ember iris
idle owl
#

But it means that other folks know someone started it, etc.

ember iris
#

SUCCESSFUL FAILURE

#

IT'S MEMORY RELATED

#

Ok, I'm going to go start a sourdough loaf and then upgrade to the beta release and see if it's there still

idle owl
#

Huzzah!

idle owl
orchid basinBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
unique roost
# tulip sleet Are you storing the encrypted password and its key on the Titano? Are you sendin...

@tulip sleet The idea is to encrypt a file on the host, then place it on the CIRCUITPY drive. That file would be decrypted by the device firmware (python) after a master password is provided by the user. That's why the Titano would be great because the password could be typed on the screen rather than be sent by a console program to the device. Once the devicce receives the password, it decrypts the file in memory, so CIRCUITPY would never show the decrypted file.

ember iris
idle owl
ember iris
#

Awesome, then I'll test it regardless soon, and I don't see anything else that stands out to me

idle owl
#

I'll approve it and merge it.

#

She probably would have merged it as-was if I hadn't said I would test it.

#

But I got in before her so she held off. 🙂

ember iris
#

Makes sense, I only noticed the added delay because I was hunting down my bug. They're right that the code would definitely work with the added delay, I just didn't know why it was there since I couldn't find a motivation in the datasheet.

idle owl
#

I've actually run into troubleshooting other things and wondering why there were delays added to things. So it was a good catch. Because it can cause issues later with something totally unrelated.

manic glacierBOT
#
urish
[adafruit/circuitpython] New comment on issue #5212: USB output only starts after main\.py finishes \(RP2040\)

So if main.py is an infinite loop, you never see the EP0_IN_BUFFER_CONTROL write?

Correct

You would probably want to watch for the execution of various routines in supervisor/shared/usb.c and tick.c. See usb_background() and usb_irq_handler() in particular

This was a very good pointer, thanks!

So here's what I'm observing:

  1. dcd_init() is called, and registers dcd_rp2040_irq() as IRQ handler for USB.
  2. USB peripheral is enabled, and then the reset sequence happens....
idle owl
#

@ember iris So you're pretty confident your resetting issue is memory related and not SCD-4x related?

ember iris
idle owl
ember iris
#

I'd say no. I'm about to update to the beta after I reply to the forum question you linked me too, then I'll be able to see if it's a memory issue that's still around.

#

So go ahead and disconnect it

idle owl
#

Right on. Will do.

#

Now I can see if Arduino works from the IDE version or if I have to do a bunch of nonsense to get the latest version from GitHub.

ember iris
#

If it the beta for circuit python solves my boards reset issue, I'll hunt down whatever issue was close between friday night's nightly build, and mondays beta and make note of the problem so we can build confidence in it's solution, but now I'm pretty confident it isn't the sensor in any way.

idle owl
#

It ran for 3:53:23 without restarting. 🙂

ember iris
#

I think that's plenty of time. I was able to make it shove a giant list into memory, and delete it, and when printing gc.mem_free() I was able to confirm that when it got super low and would normally reset was the same time the board crashed and rebooted. I *recall* (but could be wrong about that) an issue that was marked as closed regarding the call for gc free, so I'm somewhat confident the latest build is free from my issue

idle owl
#

Right ok. Makes sense.

ember iris
#

It would explain the reason the board didn't reset when I removed the sensor as well, because it didn't occupy the same ram. Additionally it explains why it was so constant with respect to the time it took to reset

idle owl
#

Exactly.

slender iron
#

@gilded cradle how's the editor coming?

idle owl
manic glacierBOT
#
todbot
[adafruit/circuitpython] Issue opened: #5216 Enable key velocity for MIDI keybeds by adding "ticks\_ms" property to keypad\.Event

The keypad module can be used for MIDI keyboards, and provide velocity sensitivity, with the addition of a millisecond timestamp to the keypad.Event object.

In a recent @jedgarpark livestream, he demo'd a MIDI keyboard using keypad. Like most keybeds, the one demo'd is implemented as two matrices: one for key down, one for key up. The intended use is key velocity, computed by measuring the time between "keypress begin" (key-up matrix 'r...

onyx hinge
#

@idle owl how do I fix this? AttributeError: 'module' object has no attribute 'SPEAKER_SHUTDOWN'
do I just need to update adafruit_macropad?

idle owl
onyx hinge
#

OK I have a local fix for now

idle owl
manic glacierBOT
idle owl
#

After insisting on making it not backwards compatible because I would fix the only instance of it! I did not. Fixed now though!

slender iron
#

@devout jolt you around and free to test funhouse dotstar status for me?

manic glacierBOT
#
jepler
[adafruit/circuitpython] Pull request opened: #5218 Fix audio sample rate on rp2040

There's a simple fix, that uses 64 bit arithmetic where necessary. But I chose to make a more complex fix, because I also noticed that it could take up to 100ms for the sample rate to be found! The new algorithm is faster (I think it runs in log(max_denominator) time, sort of like gcd) and can also give slightly more accurate results due to comparing the frequencies as floating point instead of integers—it only makes a difference in cases where the sample rate was already being met to withi...

#
jfurcean
[adafruit/circuitpython] New comment on issue #5021: MagTag Vaccination tracker crash in magtag\.network\.connect\(\) between 7alpha3 and 7alpha4

Commenting out the gc.collect() line in network.py of adafruit_circuitpython_portalbase allows the shortened example to work. The longer example still doesn't work. I think there is some interaction with gc.collect() and the network connection.

I tested this with the simple magtag example in adafruit_circuitpython_magtag and added magtag.network.connnect(). It would crash unless I also commented out the gc.collect() in add_text and set_text in __init__.py in `adafruit_...

slender iron
#

@tulip sleet @onyx hinge would you mind unassigning yourself from any 7.0.0 issues that you haven't started? that way I can know what I can pick up

manic glacierBOT
slender iron
devout jolt
manic glacierBOT
#
urish
[adafruit/circuitpython] New comment on issue #5212: USB output only starts after main\.py finishes \(RP2040\)

Also, assuming that we find a way to fix or work around the issue, there's still some time between when main.py starts running and USB Serial is up.

This means that the user is not going to see the first few prints, and more importantly, the output in case of a syntax error. I'm wondering if there is anything we can do to delay running main.py until after USB Serial is ready? Perhaps something in boot.py?

onyx hinge
#

@slender iron OK, I just entered a PR for 5092 and checked on the status of #4562, if that's not resolved I can do it next.

#

and I think that's all that's assigned to me with the 7milestone on it

slender iron
#

👍

devout jolt
slender iron
#

it is 5 dotstars right?

devout jolt
slender iron
#

that sounds right!

devout jolt
#

yep looks good to me too

onyx hinge
#

solid white in repl is intended?

slender iron
#

yup

onyx hinge
#

dons sunglasses

slender iron
#

well, it should be a dim white

devout jolt
#

thankfully they are dimmed

slender iron
#

it sets it once so it only "sticks" with neopixels and dotstars

onyx hinge
#

ok I'm relieved by them being dimmed

slender iron
#

I'm not a monster

#

😎

devout jolt
#

looks like it's drawing 10mA for all 5 dotstars (when dim white)

slender iron
#

it doesn't expect you to care about battery if in the repl

#

blinking in the repl is complicated because then you need to share the pins

manic glacierBOT
devout jolt
#

like the "pirate" build, I want to see the "tannewt is a monster" build where a loud klaxon goes off semi-periodically every 15 seconds on the piezo while in the REPL to indicate "all is well", all LEDs are on at full brightness, displayio randomly rotates, and all code statements must be prefixed with "please"

slender iron
#

yikes

#

it would be cool to beep the buzzer with status

devout jolt
#

morse code

gilded cradle
slender iron
#

kk, nice! I'm excited to try it

gilded cradle
#

🙂

manic glacierBOT
manic glacierBOT
manic glacierBOT
#
capellini
[adafruit/circuitpython] New comment on pull request #5221: Use MP\_REGISTER\_MODULE with displayio, terminalio, and fontio

I tested on my pybadge loaded with a test project using the following procedure:

  • make clean && make
  • copy firmware.uf2 to pybadge
  • ensure sure project code works
  • refactor module and temporarily comment out MP_REGISTER_MODULE line
  • make clean && make
  • copy firmware.uf2 to pybadge
  • ImportError on pybadge
  • uncomment MP_REGISTER_MODULE line
  • make clean && make
  • copy firmware.uf2 to pybadge
  • ensure sure project code works
manic glacierBOT
manic glacierBOT
tulip sleet
slender iron
#

Kk. Thanks!

manic glacierBOT
manic glacierBOT
#
dhalbert
[adafruit/circuitpython] New comment on issue #5212: USB output only starts after main\.py finishes \(RP2040\)

Yes, there is a race between main.py and USB coming up. USB is not initialized in the boot.py because you can make changes to USB there.

I do see this race when using tio, which reconnects immediately when it sees that USB is up. A single print("Hello") does not show output after a hard reset, but does after a soft-reload (ctrl-D). We did not see this as an issue before because it was not possible to manually get a connection that fast. However, tio and your simulator reconnect immed...

manic glacierBOT
#
Neradoc
[adafruit/circuitpython] Pull request opened: #5222 Make the board ID available in board and boot\_out

This implements #5105 but I am not sure of how much space it takes (for small boards), am I doing strings right ?

Additional line in boot_out.txt (there happens to be a boot.py)

Adafruit CircuitPython 7.0.0-beta.0-33-g802ca5425 on 2021-08-25; Raspberry Pi Pico with rp2040
Board ID:raspberry_pi_pico
boot.py output:

board.ID

Adafruit CircuitPython 7.0.0-beta.0-33-g802ca5425 on 2021-08-25; Raspberry Pi Pico with rp2040
>>> import board
>>> board.ID
'raspberry_pi_pic...
manic glacierBOT
#
hathach
[adafruit/circuitpython] New comment on issue #5212: USB output only starts after main\.py finishes \(RP2040\)

This sounds like a TinyUSB issue. dcd_init() shouldn't setup the interrupt handler. That should be left to the client. @hathach thoughts?

CircuitPython intercepts the interrupt so that it can queue up the USB background task. By missing the interrupt, it misses the queued work.

It is originally from raspberry pi team, I remembered we have an discussion with them before pico is released and result in API to get and set irq handler API. It shouldn't be an issue with current implemen...

manic glacierBOT
#
urish
[adafruit/circuitpython] New comment on issue #5212: USB output only starts after main\.py finishes \(RP2040\)

But I don't see anything in MicroPython that waits for serial to come up either, so I'm not sure why you don't see the problem there. It also does tud_usb_init(), and it does not output anything to USB CDC unless tud_cdc_connected() is true.

With MicroPython I do see that problem, so the default is to use the UART console and not the USB CDC while running in the simulator.

We could wait a short time for tud_cdc_connected() to be true, but we don't really know that it will ever be tr...

manic glacierBOT
manic glacierBOT
#
microDev1
[adafruit/circuitpython] Pull request review submitted: #5222 Make the board ID available in board and boot\_out

Thanks for taking this on @Neradoc. I have one comment regarding doing this as board.ID.
In my opinion board isn't a perfect place for this as it is meant for pins and there can also be a conflict with a pin named ID.

In #5051, @dhalbert suggested doing this in os.uname and that would be my preference as well.

# as a part machine 
>>> os.uname()
(sysname='rp2040', nodename='rp2040', release='7.0.0', version='7.0.0-alpha.6 2021-08-12', machine='Adafruit QT Py RP2040 (...
manic glacierBOT
#
Neradoc
[adafruit/circuitpython] New comment on pull request #5222: Make the board ID available in board and boot\_out

I chose not to add it to uname.machine because it feels wrong to have to parse it to get at the id. In this case, (no regex, promise) machine.split("(")[1].split(")")[0] ? I'd prefer the actual board ID be a clean separate field.

I don't know if machine_id is an option, it's not a C-python thing. Can we extend os.uname like that ?

An alternative mentioned in the original discussion (on circup) would be to replace uname.nodename. On Unix it's supposed to be the "network name" (though...

manic glacierBOT
manic glacierBOT
manic glacierBOT
#
dhalbert
[adafruit/circuitpython] New comment on pull request #5222: Make the board ID available in board and boot\_out

os.uname is not even present on non-Unix systems (e.g., Windows). The platform module seems to be the preferred place for an OS-independent source of information about the machine you're running on. However, there is no field there for describing the model name of the hardware, which is really what we're looking for, e.g. "Dell Optiplex 5060", or "Circuit Playground Express".

This info is extracted from the BIOS or equivalent on PC's, ...

manic glacierBOT
#
microDev1
[adafruit/circuitpython] New comment on pull request #5222: Make the board ID available in board and boot\_out

it feels wrong to have to parse it to get at the id. In this case, (no regex, promise)

I guess we can do something like Adafruit QT Py RP2040 (id: adafruit_qtpy_rp2040) with rp2040 to make it more clear.

I don't know if machine_id is an option, it's not a C-python thing. Can we extend os.uname like that ?

Good point, I totally forgot the CPython compatibility aspect there.

An alternative mentioned in the original discussion (on circup) would be to replace uname.nodename....

ember iris
#

What is the github issue tag for 'issues we need to fix before 7.0 stable'?

analog bridge
ember iris
analog bridge
#

quick q: what is the function for extracting just soc/platform i.e. MICROPY_HW_MCU_NAME like (esp32s2 or rp2040) ?

ember iris
#

So I thought I was having the same bug as in https://github.com/adafruit/circuitpython/issues/5021 using a metro exp32-s2 but suddenly I cannot force the bug in any form to occur, and I'm running a direct copy paste of the original program with no changes. Is it worth making a note that after a hard crash and a boot into safety mode, followed by pressing the reset button, I've lost the ability to reproduce the bug using the 7.0 beta?

tulip sleet