circuitpython-dev

[pastel panther]

oooh, fancy

[timber mango]

ESC[0;1;32m type stuff

[pastel panther]

right

One of these days I want to write a curses based UI to run on one of these guys

[errant grail]

@pastel panther ... one step at a time...

[pastel panther]

every time I ls /Volumes/CIRCUITPY I'm mildly nervous that it'll be gone

😱

[manic glacier]

dhalbert

[adafruit/circuitpython] New comment on issue #1327: latest master does not boot for feather\_m0\_adalogger

Following up on #1326 and #1327:

Made some CPX builds to test NeoPixels.
168e23e and after don't work: no REPL
9d91111b does not seem to build: I get this error:

In file included from ../../shared-module/usb_hid/Device.c:33:0:
../../lib/tinyusb/src/tusb.h:92:7: error: "CFG_TUD_CUSTOM_CLASS" is not defined, evaluates to 0 [-Werror=undef]
   #if CFG_TUD_CUSTOM_CLASS
       ^~~~~~~~~~~~~~~~~~~~
cc1: all warnings being treated as errors
../../py/mkrules.mk:81: recipe for target ...

[timber mango]

I find the more time that has passed between a difficult technical accomplishment, and my reexamination of it, the more likely it feels 'stable' subjectively, to me -- the more I'm apt to successfully re-navigate its ins and outs. ;)

There's no cheat for that -- no I know KungFu neo declaration for that.

The time itself is what does it.

[pastel panther]

I think I know what you mean

[timber mango]

For me, the ability to flip a switch ON or OFF (easily, without a screwdriver) has human remembrance consequences that stabilize the longer I do not touch the ON/OFF switch. ;)

If I flipped it twenty times in the past hour, I won't be able to tell you which state I left it in, without very special concentrated effort to commit the final state (ON / OFF) to human memory (to remember it).

After several hours driving in Pennsylvania on the Interstate at night, I cannot tell if I'm going uphill or downhill. ;)

[manic glacier]

bboser

[adafruit/circuitpython] Fork created: bboser/circuitpython

[solar whale]

Wow! Blinka on the Raspberry Pi is really nice! Just discovered that I can write text to my OLED display using bitmapfont.py just like I can on my feather M0/4 boards. just copied bitmapfont.py and font5x8.bin to the local directory and it works... Maybe I'm just easily impressed 😉

[pastel panther]

@solar whale no you're right, that's pretty cool

[sly wagon]

Quick WIP update on my black panther cosplay LED upgrade using circuitpython

46% of the way done. Not using soldered connections(yet) for digital pins on circuit express playground. Center animation will work after that. Ribs connection needs to be fixed too along with upper arm animation. Need to turn off green light as well

[river quest]

@sly wagon great work, we're adding this to the newsletter this week!

[sly wagon]

!!!!

That's so cool! Thank you!!! @river quest

[river quest]

email [e-mail] with any updates, photos, etc. too, we can do a blog post or whatever, circuitpython was made for this, thanks for showing the progress and more!

[sly wagon]

Will do for sure 😃 !!!

[digital slate]

I was reading through the docs for troubleshooting when the CIRCUITPY drive isn't showing up and saw this: https://learn.adafruit.com/welcome-to-circuitpython/troubleshooting#circuitpy-drive-does-not-appear-18-12

I don't have Kaspersky though and didn't see it on Ubuntu either

was wondering if there's any other known issues

I was thinking it might be the flash memory on my custom board might not be soldered down properly. I'll try resoldering it tomorrow but wondering if there's anything else to look at

[idle owl]

@digital slate There are more reasons why it wouldn't show up but most of them are software related when in the context of issues we typically run into. (Which is to say, most times it's not someone creating custom boards.) The software can become corrupted and cause issues, sometimes things will mess with the host USB controller and cause it to not show up...

Trying to think. There's an issue right now with one of the builds that causes CIRCUITPY to show up and then disappear and not show up again until you reload CircuitPython in a particular manner...

[digital slate]

yeah, was wondering if it might be a driver thing

[idle owl]

The Kaspersky thing is in the troubleshooting page because we've been seeing it so often lately, and shouldn't imply that it's the only thing that would cause it.

[digital slate]

I tested the board and the REPL shows up

so that's good

it's just the drive

[idle owl]

Which OS are you on?

[digital slate]

Windows 10 normally but like I said I tested on Ubuntu as well

[idle owl]

Neither of those should have driver issues. Hmm.

[digital slate]

it should show up without driver installs, correct?

maybe i should try reloading the bootloader

[idle owl]

Unless you installed drivers for something else that is interfering. Which can happen. As in installing drivers when you don't need them.

Yes correct. Should work without drivers.

[digital slate]

yeah, didn't install anything

esp in Ubuntu

[idle owl]

Ok. Then that's probably not it then.

[digital slate]

another data point

I have a Feather M4 and that worked fine

so it's capable of showing a drive

also when I hit reset, FEATHERBOOT shows up

[idle owl]

and you didn't change anything in your custom build that might have done something with mounting the drive?

[digital slate]

don't think so

I did build this firmwrae myself

so maybe something inthere

[idle owl]

Building it yourself shouldn't change anything like that.

[digital slate]

UF2 Bootloader v2.0.0-adafruit.7 SFHWRO
Model: Feather M4 Express
Board-ID: SAMD51J19A-FeatherM4-v0

[idle owl]

I meant actually going into the core code and altering it

before building

[digital slate]

no, this is right off of GitHub for UF2

[idle owl]

Ah ok. And what's custom about your board?

[digital slate]

not too much, I basically refactored the feather

different shape and the pins go different places

but kept everything else the same

AFAIK

[idle owl]

Got it, so that's why you're thinking the feather build should work

[digital slate]

correct

esp for getting started here, might remap pins later

but might not

[idle owl]

Thing is CircuitPython does not need the SPI flash (there are boards without it), but that build assumes it's there. So maybe you're right that the flash chip soldering isn't solid?

You said you mounted CIRCUITPY from the Feather M4, so I would agree it's not a driver or computer software issue.

[digital slate]

yeah, had some footprint issues with the part

so might not be soldered properly

question about this page: https://learn.adafruit.com/adafruit-feather-m0-express-designed-for-circuit-python-circuitpython/uf2-bootloader-details#making-your-own-uf2-45-42

[idle owl]

If your bootloader is mounting, I don't think it's a bootloader issue, but redoing the bootloader isn't a terrible step to try. I really don't think that's it though.

[digital slate]

did I miss a step and I need to convert the uf2 after compilation?

[idle owl]

Hmmm.... no. That shouldn't have been necessary.

[digital slate]

oh wait

that's the uf2

[idle owl]

Because building CircuitPython produces firmware.uf2 on its own

[digital slate]

I did the .bin from the build

[idle owl]

OH!

[digital slate]

and loaded that at location 0

so this was programmed using a JLink

[idle owl]

Wait, like using the JLink or something?

Ok, that should also work though

[digital slate]

yeah, it's the stuff I was asking about Friday

[idle owl]

but if you have bootloader, you can use the firmware.uf2 produced from your build of CircuitPython, and drag it over, I think....

[digital slate]

ah, I didn't build CP yet

[idle owl]

Loading the .bin with the JLink produces a lot more feedback if you're trying to debug something in that section

[digital slate]

I built the uf2 bootloader

[idle owl]

okay.......

[digital slate]

and then loaded the .bin to the board using a JLink

[idle owl]

How did you get the .bin?

if you didn't build it

[digital slate]

oh sorry

I mean i built the UF2 bootloader, which produced a .bin, not a .uf2

and then I loaded the .bin onto the board

which then gave me a FEATHERBOOT drive

[idle owl]

Ok. I think that means you still have to load CircuitPython if you want the CIRCUITPY drive. The bootloader is separate from CP

[digital slate]

which I dropped the .uf2 file onto

[idle owl]

Ok

you did that then

[digital slate]

(of CP)

[idle owl]

Alright. Back on the same page again

[digital slate]

yep

haha

and I am decently certain the CP is working, since i can get to the REPL via a serial terminal

[idle owl]

Right, that seems accurate.

[digital slate]

at least part of it is working

[idle owl]

Oooh question. Can you check the mount points more specifically and see if it mounted but didn't show up in the file explorer?

That has happened

As in see if the drive is there but doesn't show up on its own in the available drives in the main bit of the file explorer

[digital slate]

ah, interesting

that's like in the drive section of windows?

[idle owl]

Hmm not sure in Windows....

[digital slate]

ah

wait, wrong board plugged in

haha

[idle owl]

I would go check /Volumes/ in Mac for example

oops 😄

[digital slate]

"computer management" is the screen I was thinking of

but it's not there

should the drive have a name when it shows up in hardware manager?

[idle owl]

Ok. If it's not there, it's probably not connected. I'm pretty sure that's the most base level drive list on Windows.

um.... I'm not sure about that

[digital slate]

ok, I think I need to mess with the hardware again tomorrow

I have another board I can try this on as well

(tomorrow when I'm back in the office)

appreciate the troubleshooting help!

[idle owl]

For sure! Sorry it wasn't something simpler. At least you've got enough stuff to test with to determine it's probably not the computer. So many variables to deal with 😄

[digital slate]

yeah, it really can add up, even for seemingly simple stuff

[idle owl]

Nothing is ever simple ™

[digital slate]

haha

[idle owl]

It seriously never is though.... it's always one more thing you never expected

[digital slate]

ah, one other datapoint i just noticed as I went to go close some files

the part I have on board is the ATSAMD51J19

I believe this is similar to the feather m4

since i loaded from that

didn't noitce that though and was following tutorials that said ATSAMD51J20

might be some register differences preventing?

from when I was putting the boatloader on there

so might just retry thta with the correct part number

[idle owl]

I'm not sure. I feel like the big changer is the letter not the number... so the ATSAMD51G vs J. But I'm not sure. Datasheet might have that answer although that's a rabbit hole all its own right there.

Definitely worth retrying with the correct part though.

My Feather M4 has a J19 on it.

[digital slate]

yeah, will add that to the things I try tomorrow

thanks again!

[lone sandal]

arofarn

[adafruit/awesome-circuitpython] New star added

[lone sandal]

CedarGroveStudios

[adafruit/awesome-circuitpython] New star added

[manic glacier]

ghhann

[adafruit/circuitpython] Issue opened: #1328 Cannot upload ardunio sketch if 3\.11 executing

Arduino: 1.8.7 (Windows Store 1.8.15.0) (Windows 10), Board: "Adafruit Feather M0 Express"

C:\Program Files\WindowsApps\ArduinoLLC.ArduinoIDE_1.8.15.0_x86__mdqgnx93n4wtt\arduino-builder -dump-prefs -logger=machine -hardware C:\Program Files\WindowsApps\ArduinoLLC.ArduinoIDE_1.8.15.0_x86__mdqgnx93n4wtt\hardware -hardware C:\Users\Glenn\Documents\ArduinoData\packages -hardware C:\Users\Glenn\Documents\Arduino\hardware -tools C:\Program Files\WindowsApps\ArduinoLLC.ArduinoIDE_1.8.15.0_x86__md...

[manic glacier]

deshipu

[adafruit/circuitpython] New comment on issue #1328: Cannot upload ardunio sketch if 3\.11 executing

I think that you need to press reset twice to get the board into the bootloader mode.

[manic glacier]

caternuson

[adafruit/circuitpython] Issue closed: #1299 Driver needed for MPL115A2\.

caternuson

[adafruit/circuitpython] New comment on issue #1299: Driver needed for MPL115A2\.

Closing.

Code added to repo:
https://github.com/adafruit/Adafruit_CircuitPython_MPL115A2/pull/1
Repo added to bundle:
https://github.com/adafruit/Adafruit_CircuitPython_Bundle/pull/111

[exotic pumice]

It seems that when I try to feed DFLL into GCLK0 my program hangs

[manic glacier]

tannewt

[adafruit/circuitpython] New comment on issue #1322: Rewriting MCP230808 I2C Circuitpython library to use SPI for MCP23S08 chip

This should go in the community bundle since its not carried by Adafruit.

[manic glacier]

tannewt

[adafruit/circuitpython] New comment on pull request #1323: shared\-module/os: Fix os\.mkdir\('a/b'\)

I'd copy what CPython does on a UNIX style system.

[slender iron]

@digital slate Sounds like the build isn't configured correctly for the flash chip. Over serial do:

import os
os.listdir("/")```

if it raises an error there is a spi flash issue (not a host os issue)

<@&356864093652516868> Here are the notes for tomorrow's meeting: https://docs.google.com/document/d/1z3-F_GTPGvw5fETP39sFrMuKw3qlRyQftPCn9bAJaRU/edit?usp=sharing

[exotic pumice]

I should be able to make it to circuitpython weekly this week 😀 It's a national holiday in Canada tomorrow so I don't have any school

[manic glacier]

ngmacha

[adafruit/circuitpython] New comment on issue #1322: Rewriting MCP230808 I2C Circuitpython library to use SPI for MCP23S08 chip

Thank you - but could you kindly clarify: is that something that I'll have to take actions? Is this a wrong place to raise this issue?..

[covert oxide]

Hi, I have a question, can any port be set to double-precision floats? I tried setting MICROPY_FLOAT_IMPL to MICROPY_FLOAT_IMPL_DOUBLE in mpconfigport.h but I see issues "/py/formatfloat.c:94:5: error: floating constant exceeds range of 'float' " , do you know if there is something else I need to change?

[raven canopy]

clearing the FRAM breadboard, re-bagging the breakouts, and realize i have 2 of the I2C version. oddly enough, i have no recollection of ordering two, nor do i remember soldering headers on both... good thing i don't spend much time on ebay. 😆

[manic glacier]

sajattack

[adafruit/circuitpython] New comment on issue #1322: Rewriting MCP230808 I2C Circuitpython library to use SPI for MCP23S08 chip

@ngmacha https://github.com/adafruit/CircuitPython_Community_Bundle

[digital slate]

yep, looks like there's something wonky @slender iron

Nothing there

is there a default bootloader .bin I can try pushing to the chip?

I'll definitely still want to build it

[tidal kiln]

@raven canopy maybe you were thinking you'd test multiple devices on the I2C bus?

[slender iron]

Why bootloader? Seems like that is working. This is a circuitpython issue

[raven canopy]

@tidal kiln we'll go with that. since it reduces any cases of brain == absent. 😄

[slender iron]

The bootloader’s drive is fake, not the spi flash

[digital slate]

oo

got it got it

yeah bootloader does seem to work

it pops back into featherboot whenever I doubleclick the reset

[slender iron]

You’ll need to configure your board config for the spi flash device

[digital slate]

and that shows up fine on Windows and Linux

gotcha

[slender iron]

Yup that’s all good then

[digital slate]

just thought I was copying things exactly

hence my confusion

like same parts AFAIK on the Feather M4

[slender iron]

It checks the Jedec id on startup

Qspi too?

[digital slate]

yeah

let me go reslder the pins, they look shaky

maybe that's it

[slender iron]

Hrm. Then it could be an assembly issue. But I’d triple check the pin mapping and part number

[digital slate]

ok, will do

what should I see when I do that OS list command?

[slender iron]

Boot_out.txt at least

[manic glacier]

ladyada

[adafruit/circuitpython] New comment on issue #1322: Rewriting MCP230808 I2C Circuitpython library to use SPI for MCP23S08 chip

@ngmacha hiya, the issue can stay here. we don't have a plan to write this driver ourselves. if you plan to, or someone else can write it - we can include it in the bundle!

[umbral dagger]

My talk on CP at PyCon Canada went well. Packed room, lots of interest, good questions, plenty of followup discussions after the talk.

[idle owl]

@umbral dagger That's great! 🎉

[umbral dagger]

I believe it was recorded.

[tulip sleet]

@umbral dagger Drop a link to the video when you know what it is. I looked but couldn't find it.

[manic glacier]

notro

[adafruit/circuitpython] Pull request opened: #1329 Support OSError subclasses and attributes

This adds support for the OSError subclasses listed in the Python docs:

      +-- OSError
      |    +-- BlockingIOError
      |    +-- ChildProcessError
      |    +-- ConnectionError
      |    |    +-- BrokenPipeError
      |    |    +-- ConnectionAbortedError
      |    |    +-- ConnectionRefusedError
      |    |    +-- ConnectionResetError
      |    +-- FileExistsError
      |    +-- FileNotFoundError
      |    +-- I...

[umbral dagger]

@tulip sleet They don't seem to be posted yet. I'd expect later this week. I'll be sure to post the link.

[manic glacier]

notro

[adafruit/circuitpython] New comment on pull request #1329: Support OSError subclasses and attributes

Forgot to mention flash usage:

Base:
58368 bytes free in flash out of 499712 bytes ( 488.0 kb ).

OSError patch: Use 2468 bytes
55900 bytes free in flash out of 499712 bytes ( 488.0 kb ).

uerrno patch: Use 320 bytes
55580 bytes free in flash out of 499712 bytes ( 488.0 kb ).

[digital slate]

@slender iron after solder touchup on the QSPI, looks like some drives show up now

does that part get touched in the bootloader install?

[slender iron]

great!

[digital slate]

like should I reload that

[slender iron]

nope, its ignored

[digital slate]

but the drives are empty, as is the os.listdir

[slender iron]

no boot_out.txt?

[digital slate]

but some progress is better than none

nope

tried reloading the latest CP build as well

[idle owl]

@digital slate Great job!

[slender iron]

@digital slate can you load a file from the comp and have it stay?

[manic glacier]

notro

[adafruit/circuitpython] New comment on pull request #1329: Support OSError subclasses and attributes

Version 2

• Fix viper_misc test

[digital slate]

it's still not showing up as a CIRCUITPY drive, to be clear

let me try dropping something onto there in a bit

[slender iron]

oh, I thought it was

[digital slate]

no, they're blank drives, there are two of them

it's weird

[slender iron]

ah, that sounds suspect. you can try to reformat them from the repl

storage.erase_filesystem()```

[raven canopy]

<@&356864093652516868> and any/all who wish to participate: weekly meeting in about 40 minutes. Notes are here: https://docs.google.com/document/d/1z3-F_GTPGvw5fETP39sFrMuKw3qlRyQftPCn9bAJaRU/edit?usp=sharing

[fluid helm]

Not gonna make it today folks, sorry

[hazy gust]

I'm new to all things. sorry. When you use REPL, you need to enter each line of code in? You cant, like, set a breakpoint and troubleshoot 1 part/section?

[stuck elbow]

No, it's a repl, not a debugger. There is some work ongoing to provide a debugger as well, but I think it has stalled a bit.

[raven canopy]

@hazy gust unfortunately, no, we don't have breakpoint-style debugging. like @stuck elbow stated, it has been brought up and discussed: https://github.com/adafruit/circuitpython/issues/298

[hazy gust]

cool, no worries, thank you.

[exotic pumice]

my discord is saying RTC Disconnected, not sure what that means

[stuck elbow]

real-time chat, I guess

[raven canopy]

@exotic pumice mine did too when i first tried to connect. try re-connecting.

[gusty kiln]

yeah, seems like discord is a little glitchy today.

[strange pumice]

With the M0 Express Feather is it possible to tell when it is connected to a USB power source?

[gusty kiln]

brb.

[inland tusk]

goood day everybody

[gusty kiln]

(back.)

[fierce girder]

Scott: Travelling so I'm just lurking today

[marble hornet]

anyone else have no audio ?

[timber mango]

@strange pumice yes and no.

[marble hornet]

nm

[tulip sleet]

@inland tusk we're hearing background noise ; could you mute?

[strange pumice]

@timber mango - The only way I can think of is to have a battery connected and watch the voltage change. Obviously this would be unreliable if the battery were fully charged. What ways do work?

[timber mango]

@strange pumice let's go over to #general-tech and discuss it.

[marble hornet]

Tannewt: i'm just lurking, b/c only been working on my project

[idle owl]

@hazy gust We're in a meeting right now, FYI - may be a bit for a response.

Oh I wasn't scrolled. Someone already responded.

[exotic pumice]

I added a new PR this morning in libraries 😃

[tulip sleet]

@strange pumice you can tell if there's a usb serial connection: supervisor.runtime.serial_connected

[idle owl]

@exotic pumice We had already pulled the stats before you put that in. You'll be included in next week's.

[exotic pumice]

yep

[raven canopy]

@strange pumice note to dan's suggestion: serial_connected is not totally dynamic. it will update when a USB connection is made, but not when a disconnect happens.

[idle owl]

CircuitPython Library Tracking Issue: https://github.com/adafruit/circuitpython/issues/1246

[strange pumice]

Good to know, thank you both @tulip sleet and @raven canopy

[marble hornet]

can one hug report even if just lurking?

[raven canopy]

bug...fixed!

[marble hornet]

thanks!

[idle owl]

@marble hornet Feel free to type it in now, you'll be next 😃

[marble hornet]

hugs to: @tulip sleet, @solar whale , @tidal kiln , @slender iron , and @raven canopy for always being so kind and willing to help 😊

also for helping debug

[pastel panther]

stuff in notes plz

[slender iron]

To kattni, scott, cgrover, nis, sommersoft and everyone else who helped me troubleshoot the issues with the board who shall not be named and talked me off the ledge when things went more and more pear-shaped.
To cgrover for a reminder that styrene is a thing that exists; it made my bezel making waay easier

[marble hornet]

who said that ^^^?

[pastel panther]

meeee

[marble hornet]

👌 tnx, just didn't catch name

i'm good thanks

[tulip sleet]

@gusty kiln Drew Fustini might have some intelligence on gpiod packages and their eventual appearance in Raspbian.

[idle owl]

Yeah Drew's in on the issue for it.

[fierce girder]

Yeah

I'll think about the packaging issue

We do have it on the Debian image for BeagleBone

But yeah raspbian is the more important platform

[gusty kiln]

i kind of lean towards "use it if available, otherwise fall back"

[fierce girder]

@errant grail good to hear 😀

@gusty kiln that makes sense

[gusty kiln]

that way we could tell people how to install but not make it a prereq for blinka on the pi.

[raven canopy]

studio builds...such fun.

[fierce girder]

It will be ahwile before /sys/class/gpio dissappears

[pastel panther]

so fun. much wiring

[fierce girder]

This is good work for the future

Yeah

[errant grail]

The Trinket M0 meter tester.

[marble hornet]

NEATO man

[tidal kiln]

@errant grail do you use eagle or kicad?

[errant grail]

KiCAD 5.0

The Feather Wing version of the meter tester.

[tidal kiln]

how'd you make the trinket footprint?

[errant grail]

Manually with the footprint editor.

[tidal kiln]

OLED sneeze guard?

[errant grail]

Works for fingerprints, too. I have a very curious grandkid...

[manic glacier]

notro

[adafruit/circuitpython] New comment on pull request #1329: Support OSError subclasses and attributes

Version 3

• Really fix the test including the expected result file

[marble hornet]

@ leaving the protective film on forever

[neat folio]

@marble hornet I believe the technical term is cheese wrapper, or so I have been told. 😉

[marble hornet]

ohh is that what's use to go between cheese levels ? like the bluefruit?

😉 😅

[neat folio]

hmmm.... meta... now my brain has to work...

[exotic pumice]

https://github.com/atsamd-rs/atsamd

[pastel panther]

Stuff in notes again, please. pix incoming

[idle owl]

From @pastel panther:
last week: Received and assembled the new boards, didn't work, couldn't figure out issues in time for a new board rev :(((((
Scrapped the first board and assembled a second one that is working great though the boost converter needs rework before assembly
Learned a lesson: never. ever trust someone else's footprints.
Resolved to have a single example of the blinky Cyborg Raccoon for display at Designer Con instead of having some for sale :|
Started assembly of a base, doing some woodwork to build in the TFT and made an initial bezel out of styrene.

This week:
Finish up the rest of the base, prep and paint.
Work on more animations for the TFT and finalize coding for TFT and LED animations.
Pack up the car and drive to Anaheim for Designer Con! http://www.designercon.com/
(we're in booth 1049 for anyone who wants to stop by)

[marble hornet]

@pastel panther what screen is that ?

[solar whale]

Cool! BluefruitSPI allowing control of CPX neopixels from iPhone in CircuitPython

[pastel panther]

@marble hornet https://www.aliexpress.com/item/maithoga-0-9-inch-13PIN-262K-SPI-TFT-LCD-Color-Display-Screen-ST7735S-Drive-IC-80/32841634717.html (via hot tip from @stuck elbow )

[marble hornet]

does it work with the st7735r lib?

[pastel panther]

ya

with some tweaks

[marble hornet]

ooh

[stuck elbow]

by the way, it would be nice to add those tweaks to the library at some point

[pastel panther]

Definitely will.

[marble hornet]

as another class?

[stuck elbow]

possibly

[pastel panther]

No, just need some offsets

[solar whale]

Still waiting for my particles too 😦

[pastel panther]

And some register munging to fix the rotation

[marble hornet]

oh, is that the same as on : https://www.adafruit.com/product/3321 ?

[idle owl]

@raven canopy Nice! Thank you!

[pastel panther]

not sure

looks the same

[tidal kiln]

IN THE WEEDS

• still need flash eraser UF2s?

[pastel panther]

ahahahahha

again!

[raven canopy]

yeah...Google can be mean. 😦

[stuck elbow]

it is the same

[gusty kiln]

blobfish: does what says on tin.

[neat folio]

after the weeds: talk about scheduling a get together for learning?

[inland tusk]

It does look like that in the water

[slender iron]

https://github.com/adafruit/samd-peripherals/blob/master/samd/samd51/clocks.c#L106

[pastel panther]

the '51 errata is a graveyard of broken dreams

[raven canopy]

https://github.com/adafruit/circuitpython/issues/990

[exotic pumice]

XD 32 pages of errata

[pastel panther]

wonder if you could patch uf2-samd to jump to some different offset with some reset tap pattern?

[slender iron]

@pastel panther I've been brainstorming that too. It could just set special memory values

[marble hornet]

is there a try and except in c?

[pastel panther]

also that

[tulip sleet]

https://github.com/adafruit/circuitpython/issues/990

[raven canopy]

@tulip sleet i posted it ... or that. 😄

[marble hornet]

what about the capacitive pins?

[pastel panther]

doubletapping with tweezers is also hard

[idle owl]

There's no D13 on the Trellis M4.

[pastel panther]

admin build could be cool?

(comes with free footgun)

(foot kevlar vest?)

[raven canopy]

(foot moat?)

[exotic pumice]

@raven canopy, the samd51 version is on my fork here if you're curious https://github.com/sajattack/atsamd/tree/samd51

[raven canopy]

@exotic pumice will check it out! make no promises i can help...my smarts have limits. 😄

[gusty kiln]

real popular in webdev land

[neat folio]

"polite" suggestions (replaces interrupts after being nudged) 😉

[solar whale]

nudges - not interrupts

[exotic pumice]

I just noticed, I think I typed exclamation marks in my commit messages and it got replaced with my previous command

rust has interrupts 😛

[raven canopy]

so...MQTT-lite? no living history, just last events...

[inland tusk]

@slender iron why can you not implement queues like the bus module.

[neat folio]

so the t-shirt is labelled "pre-emption DE-nied" with an image of the towering 7 foot basketballer between you and rim... and you have have to choose between walking away with nothing, or looking at message queues and saying... "I could work with this" 😃

[pastel panther]

gotta run everyone! 👋

[gusty kiln]

so i wanted something like interrupts when i was doing a project with LoRa + some other stuff a while back.

message queues feel like interface scaffolding that people wind up kind of naively re-implementing when they're polling across a bunch of stuff.

i dig it, personally.

[errant grail]

Have to hit the road, folks. Great meeting today!

[marble hornet]

better cues than none, as someone who learned it in school the cues sound great as a start

[raven canopy]

typedef struct {
    uintxx_t message[n]; // message; what's the cutoff size?
    timestamp; // systick 
} msg_queue;

top-of-the-noggin, less than stellar psuedo...

[marble hornet]

high school^^

could i ask for some advice, python wise?

after after the weeds

[gusty kiln]

yeah, fair 'nuff.

off -> lunch. later all.

[idle owl]

@slender iron Is there any reason not to do a release on this lib? https://github.com/adafruit/Adafruit_CircuitPython_CPython

[solar whale]

not annoyed at all -- but I also have to run -- Have a good day all!

[raven canopy]

same here. gotta figure out lunch. i may drop back in.

[tidal kiln]

https://github.com/adafruit/circuitpython/issues

[manic glacier]

DakotahRedstone

[adafruit/circuitpython] Issue opened: #1330 call for the creation of the first "LEARN SESSION"

a "LEARN SESSION" being a a live session analog to a LEARN GUIDE
I would like for the first session or series of sessions to be about using GitHub to do code reviews.
If I was dreaming big, this might develop into a recording series like a "coding workshop", since I think one could learn to code from doing reviews at the start.

[manic glacier]

caternuson

[adafruit/circuitpython] New comment on issue #1330: call for the creation of the first "LEARN SESSION"

For reference, adding a link to the Learn Guide that this would augment:
https://learn.adafruit.com/contribute-to-circuitpython-with-git-and-github/overview

[exotic pumice]

what speed does this achieve? 1.92 MHz? https://github.com/adafruit/samd-peripherals/blob/f20fcf642b5654ee68d7d551ea7db39716ef83bf/samd/samd51/clocks.c#L96

I'm not quite sure what the reference clock is at

I calculated based on 32KHz

[slender iron]

<@&356864093652516868> Here is the recording from today's meeting: https://youtu.be/FPqeLzMAFvA

[slender iron]

@tulip sleet debug works for me

[manic glacier]

3rc0

[adafruit/circuitpython] Fork created: 3rc0/circuitpython

[slender iron]

@tulip sleet make sure you have the latest tiny usb

[tulip sleet]

halbert@salmonx:~/repos/adafruit/circuitpython/lib/tinyusb$ git status
HEAD detached at 299a2f12
nothing to commit, working tree clean
halbert@salmonx:~/repos/adafruit/circuitpython/lib/tinyusb$ git remote -v
origin    https://github.com/tannewt/tinyusb.git (fetch)
origin    https://github.com/tannewt/tinyusb.git (push)

i'll try it on a mac

[exotic pumice]

why are clocks set to track dpll0 before dpll0 is initialized?

I'm having a hang

[tulip sleet]

that might be a mistake, but dpll0 might have been left over from the bootloader??

[exotic pumice]

https://github.com/adafruit/samd-peripherals/blob/f20fcf642b5654ee68d7d551ea7db39716ef83bf/samd/samd51/clocks.c#L120

I'm gonna try doing the initialization first and see what happens

[meager fog]

@tulip sleet hihi

[tulip sleet]

hi

[meager fog]

if we add a new UF2 bootloader def to our repo
what is the thing we do to make it generate the release files, do you do that manually

[tulip sleet]

i tag it and then build by hand and upload them

[meager fog]

oi

[tulip sleet]

yeah, it should be travis-ized

[meager fog]

ok - thanx - we have an incoming PR soon

should we have a shell script that at least iterates over everything to build?

i could probably write that

[tulip sleet]

there's a makefile target: drop-pkg or something like that. It's at the end of the makefile

[meager fog]

ooh ok thats not too bad then - ok thanx

[tulip sleet]

yeah, it doesn't take long

[sly falcon]

Howdy all...I've poking today and wanted to make sure I'm not missing anything - has anyone ever gotten the CircuitPython adafruit_rgb_display library working on the Hallowing?

[tulip sleet]

@slender iron HEAD of master works on Mac, does not work on Linux or Windows 10. On Windows 10, gets "USB device not recognized". I started with a completely fresh clone of adafruit/circuitpython, just to make sure there was no issue with tinyusb submodule

@slender iron does not always work on Mac. If I plug it in fresh, sometimes it doesn't enumerate and neopixel stays dark. If I then press reset, it does come up.

reset button on Linux doesn't help

[slender iron]

I was testing with arch linux

[tulip sleet]

try on your windows box - i assume you're out

[slender iron]

But I give it boot time before connecting usb. Ya at the gym now

Will try later. Probably a race condition of some sort

[tulip sleet]

we're going to the movies but I'll be back later in the evening like after 6:30-7 your time

[sly falcon]

w00t! Nevermind - I got it working...so stoked! Thanks to all of you wonderful folks behind CP 😄

[exotic pumice]

I seem to be getting stuck waiting for dpll status to be ready

maybe a quick sanity check is in order

while (!(OSCCTRL->Dpll[0].DPLLSTATUS.bit.LOCK || OSCCTRL->Dpll[0].DPLLSTATUS.bit.CLKRDY)) {}

is the same as

while oscctrl.dpllstatus0.read().lock().bit_is_clear() ||
        oscctrl.dpllstatus0.read().clkrdy().bit_is_clear() {}
}

oh wait no it's not

tricky boolean logic

still not happy

[manic glacier]

bboser

[adafruit/circuitpython] Fork created: bboser/circuitpython

[ruby lake]

hm, is there a win7 usb serial driver exe that includes the trellis m4?

[slender iron]

@tulip sleet I'm back working now if you get back

[inland tusk]

@idle owl is there a guide on how to get started with github for adafruit?

Also I would like to know if there is a template for creating libraries for drivers?

[timber mango]

https://git-scm.com/book/en/v2 for advanced stuff in git

You can clone a git repository on your hard disk -- to another directory on your hard disk.

Then you can push changes from the clone, back to the original.

In that way, your edits are not exposed to the public (as they would be on github.com).

[meager fog]

@slender iron heya for tinyusb do you get enumeration on samd51?

[slender iron]

@meager fog I do when I power it, start it and then connect usb. Its flaky on samd21 if I connect immediately

I think it may be because don't wait for the enable sync. Adding it now to see

[solar whale]

@inland tusk see https://learn.adafruit.com/creating-and-sharing-a-circuitpython-library

[idle owl]

@inland tusk That is for creating libraries. This is for general Git/GitHub workflow: https://learn.adafruit.com/contribute-to-circuitpython-with-git-and-github/

[meager fog]

@slender iron kk just checkin'!

[slender iron]

that wasn't it. so there are startup gremlins I gotta find

[meager fog]

~its the clocks~

lol

[slender iron]

hopefully not. I didn't change those 😃

[exotic pumice]

its always clocks

[slender iron]

it might be.....

[exotic pumice]

I have no idea, I'm just venting my own frustration

[slender iron]

its not a bad intuition 😃

[tulip sleet]

@slender iron I'm back but have a phone call to make for a while.

[slender iron]

k np, just poking at it. I can only replicate when plugging it in

[meager fog]

@slender iron k i @'d ya on basecamp if you can reply before you zzz

[slender iron]

lemme look

@meager fog just replied

[inland tusk]

@idle owl It occured to me that when you upgrade a brand new cp express, you need to replace certain libraries for the demo to work. Could you include a list so that updater could know what libraries to extract from the new bundle.

[meager fog]

@slender iron thanks - back to routing for meeeee

[ruby lake]

hm, feather m4 shows up, trellis m4 does not.

[pastel panther]

same build?

[ruby lake]

does the adafruit windows drivers know the neotrellis m4 vid/pid?

[timber mango]

I did not have to tell Linux about Trellis M4.

[exotic pumice]

I set my d13 to be !(dpll0->clkrdy) and it's always on 😢

[ruby lake]

linux is better at resolving usb devices 😉 win7 is a bit old school

[timber mango]

I seem to recall I had to do something unusual when I first got the Trellis. There's a new .UF2 bootloader update out, and is a required update.

[ruby lake]

yeah, looking at Adafruit_usbser.inf there is no USB\VID & PID entry for the trellis m4

[timber mango]

https://learn.adafruit.com/adafruit-neotrellis-m4/update-bootloader

(mine needed the update; I got mine Nov 1st or thereabouts)

[ruby lake]

right, that deals with the bootloader not being marked protected

this is getting the computer to see the trellis at all

[tulip sleet]

@ruby lake windows 7 or 10?

[ruby lake]

7

[tulip sleet]

sorry, yes, we need to update the driver package for it. I'll consult about it.

[ruby lake]

It probably needs lines like "%FEATHER_M0_RADIOFRUIT% CircuitPython (8025:00)"=DriverInstall, USB\VID_239A&PID_8025&MI_00

tuned for the trellis m4

[tulip sleet]

right, you could make your own, but it won't take that long to update. I'll try to do it tonight or by sometime tomorrow.

[ruby lake]

no huge rush, I'm about to zonk for the evening..but thanks 😉

[tulip sleet]

ok, I won't stay up late for it 😃

[slender iron]

@tulip sleet can you see what beagle shows?

I can replicate it sometimes but its not obvious where the code is stuck

[tulip sleet]

i'll get a trace

[slender iron]

thanks

I tried to get a backtrace but the stack was corrupt

[exotic pumice]

can we get a sad clock emoji?

[tulip sleet]

@slender iron here are two traces from plugging in the M4 metro: once it worked, once it didn't

[slender iron]

excellent!

the failed one looks like it started working

and beagle would be better for this too since it shows the packet level

[tulip sleet]

the failed one just hung. I'll break out the beagle

[slender iron]

it could have hung on the set configuration

bonus points for getting it to hang with jlink tracing

I gotta take a break for now though

frustrating to have something so picky

[tulip sleet]

i'll post something here when I can get it to fail

[slender iron]

thanks!

[tulip sleet]

@slender iron

^C
Program received signal SIGTRAP, Trace/breakpoint trap.
0x0001057c in maybe_transfer_complete () at ../../lib/tinyusb/src/portable/microchip/samd21/dcd.c:283
283            if (maybe_handle_setup_packet()) {
(gdb) bt
#0  0x0001057c in maybe_transfer_complete () at ../../lib/tinyusb/src/portable/microchip/samd21/dcd.c:283
#1  USB_Handler () at ../../lib/tinyusb/src/portable/microchip/samd21/dcd.c:338
#2  <signal handler called>
#3  0x00002174 in __udivsi3 ()
#4  0x0002d806 in color_brightness (color=12288, brightness=<optimized out>)
    at ../../supervisor/shared/rgb_led_status.c:189
#5  0x0000d8c4 in tick_rgb_status_animation (status=<synthetic pointer>)
    at ../../supervisor/shared/rgb_led_status.c:267
#6  run_code_py (safe_mode=NO_SAFE_MODE) at ../../main.c:278
#7  main () at ../../main.c:432

Failed the first time I tried it. I should go to sleep but that backtrace looks promising. Getting datacenter to work on Ubuntu 18.04 required adding a bunch of old .so's to their private lib dir. They've had months to update their Linux version for 18.04 and haven't done it yet.

[slender iron]

ya, that doesn't surprise me. Thanks!

[tulip sleet]

many minutes later now, bt is just like #3 down. I assume the USB went dead.

[slender iron]

ya, I think its a concurrency issue. maybe with the fifo

anyway, back to my break. thanks again!

[tulip sleet]

sure, gn!

that backtrace was a lucky break. Repeating the whole thing now and I'm just backtracing in the pulsing loop

[meager fog]

@slender iron l@@k

[pastel panther]

ooooh

[slender iron]

nice! very excited

@meager fog is there a fuse on the trellis? I think I'm triggering it

[meager fog]

yes there is

trips at 1A

[slender iron]

will all pixels at full brightness trip it?

[meager fog]

yep

z'not a lamp 😃

[slender iron]

and what happens when its tripped?

[meager fog]

the whole thing will shut off

no power

for about 30 sec

until the fuse resets

cuts the UBS 5V

[slender iron]

hrm, i'm seeing them stay on but get dimmer and yellower

and the mcu becomes inaccessible

[meager fog]

ooh

that iz a brownout

the voltage sags a ton

ive done that

[slender iron]

ya, thats what it seems like

[meager fog]

i dont know if we have BOD fuse on

i think the samd51 had some bug so we didnt

[slender iron]

I think we do

ah

welp, I'm in there now

[meager fog]

yeah congrats!!!

🎈 🎊

[slender iron]

somehow my usb change causes us to send all white 🤦

[meager fog]

rad

bad timing data can do that...but thats assuming you're sending data

[slender iron]

ya, thats a suspect

overwriting the wrong memory is too

[meager fog]

heck yeah chapter 9

[slender iron]

kinda hard to debug until I can recover from the brownout

[meager fog]

you could hot air the shifter off

its easy to replace

its the same chip as an itsy m4

[slender iron]

nah, ideally I'll make it recoverable

first, when we reset the board, set them to off

[meager fog]

its possible..

the SK's tend to come up off

unlike previous pixels

so if you depower, they should not light up

if that makes sense

like, its definiately something sending data to cause it to turn white

[slender iron]

right, but ideally we can recover from the brownout

[meager fog]

hmmm not sure how

[slender iron]

if we manage to run the start of circuitpython before things get worse

[meager fog]

the pixels are putting it into brownout, you have ot shut em off

its possible! i think?

[slender iron]

eh, worth a try

[meager fog]

not sure tho

[slender iron]

the brownout isn't immediate so resetting with ctrl-c could save you too

as long as it shuts them off for you

[meager fog]

oh right yeah

if you're already connected

[slender iron]

welp, it does it when I try and reset them at the start

[meager fog]

yay!

wait, misread. not yay

[slender iron]

totally a clock thing 😕

[meager fog]

its always clocks

https://www.youtube.com/watch?v=d020hcWA_Wg

💤 for me

[slender iron]

night!

[tough flax]

Hey @slender iron can I get a pointer to the library and/or board def for the M0+BLE? I may have time to test it tomorrow

[manic glacier]

tannewt

[adafruit/circuitpython] New comment on issue #1326: Trellis M4 latest build causing issue with NeoPixel code

Thank you for the detailed report. I was able to reproduce this on my end. I believe @jerryneedell is right that the bad failure mode is a result of overdrawing the usb line with all white pixels.

The question is, why the white pixels? Well, I haven't figured it out. It appears that the neopixel clocking is slightly slower than in d08747d. The zero high time is 0.5 us instead of 0.38 us which is enough to be a one. I get these numbers on a Metro M4 without neopixels attached so it isn't a ...

[solar whale]

@tough flax accordting to this Guide https://learn.adafruit.com/munny-lamp/code-with-circuitpython - use the feather_m0_adalogger build. I did this and it worked fine. I think the only difference from the feather_m0_basic is the definition of some of the pines used on the adalogger -- I did not look to closely to see if they are actually used but as I noted, it worked for me 😉 Looking at the code, it is not clear to me why the feather_m0_basic build would not work as well. I'll test it later.

there are exaple scripts here for testing the board https://github.com/adafruit/Adafruit_CircuitPython_BluefruitSPI/tree/master/examples

[tulip sleet]

@slender iron I'm really sorry - the USB traces I posted last night were for Metro M0, not an M4 board, despite the filenames.

[manic glacier]

dhalbert

[adafruit/circuitpython] New comment on issue #1326: Trellis M4 latest build causing issue with NeoPixel code

I wonder if the instruction/data cache is on at the right time? Maybe that got lost in some refactoring? It gets turned on in the beginning, and also gets turned off/on during some external flash operations.

[neon briar]

I'm trying to write to the SD Card on the Adalogger with CircuitPython from a Feather M4 Express. The docs don't seem to directly address using Adalogger with this setup, so I'm trying different options: https://learn.adafruit.com/adafruit-adalogger-featherwing/circuitpython#initialize-and-mount-sd-card-filesystem-3-4

Trying this:

cs = digitalio.DigitalInOut(board.D0)
sdcard = adafruit_sdcard.SDCard(spi, cs) ```
produces:
```Traceback (most recent call last):
  File "main.py", line 19, in <module>
  File "adafruit_sdcard.py", line 113, in __init__
  File "adafruit_sdcard.py", line 141, in _init_card
OSError: no SD card```

There's an 32GB micro SD card in the Adalogger. I looked at it on a Mac and it appears to be formatted.

Replacing board.D0 with board.SD_CS won't work because that attribute doesn't exist:

Traceback (most recent call last):
  File "main.py", line 18, in <module>
AttributeError: 'module' object has no attribute 'SD_CS'```

These are the attributes of board:

['A0', 'A1', 'A2', 'A3', 'A4', 'A5', 'SCK', 'MOSI', 'MISO', 'D0', 'RX', 'D1', 'TX', 'SDA', 'SCL', 'D4', 'D5', 'D6', 'D9', 'D10', 'D11', 'D12', 'D13', 'NEOPIXEL', 'VOLTAGE_MONITOR', 'I2C', 'SPI', 'UART']```

What's the correct setup to write to the Adalogger SD Card from the Feather M4 Express?

[tidal kiln]

@neon briar try D10

cs = digitalio.DigitalInOut(board.D10)

[prime flower]

TinyLoraWAN moving into a circuitpython library 😃

[meager fog]

@prime flower heya are you around? plz ping me on slack 😃

[tulip sleet]

@ruby lake new adafruit win7 drivers package with Trellis M4: https://github.com/adafruit/Adafruit_Windows_Drivers/releases/tag/2.3.3

[ruby lake]

@tulip sleet Thanks, will check it out when I get home

[river quest]

relevant #LEEK

nrf52840 bluefruit Feather

[pastel panther]

(Express?)

methinks I see qspi flash and neopixel

[river quest]

@pastel panther yep, both 😃

[digital slate]

Lookin good!

hey @slender iron I saw your post over on the CE forum about QSPI configuration

this page: https://github.com/adafruit/circuitpython/blob/master/supervisor/shared/external_flash/devices.h

[slender iron]

yup yup

[digital slate]

this is the part I have on board

https://www.digikey.com/product-detail/en/winbond-electronics/W25Q16JVSSIQ/W25Q16JVSSIQ-ND/6193769

[slender iron]

looks like its W25Q16JV_IQ in the list

[digital slate]

yep, saw that one

anyway, off to build, will let you know how it goes!

[slender iron]

kk!

[digital slate]

thanks all for the help

[slender iron]

np

[prime flower]

Does anyone using VSCode encounter it "locking up"/unresponsive window errors when editing files on the /circuitpy/ volume?

[pastel panther]

I can't say I have

I'm on a mac fwiw

[prime flower]

same

[slender iron]

@tulip sleet the disassembly of neopixel_write looks very different between the two builds

welp, the debug build was switched from -Os to -O0 which would do it

!!!! thats for all builds

[neon briar]

Thanks @tidal kiln board.D10 works. 👍

[slender iron]

(╯°□°）╯︵ ┻━┻

[tulip sleet]

well that was a rathole

[slender iron]

you have no idea

[neon briar]

@tidal kiln was there a schematic you looked at to find that? Or source code?

[idle owl]

@slender iron oi...

[manic glacier]

tannewt

[adafruit/circuitpython] New branch created: tannewt\-patch\-4

tannewt

[circuitpython:tannewt\-patch\-4] 1 new commit

ed9db80 Switch SAMD51 back to -Os - tannewt

[exotic pumice]

@neon briar, the learn guide here says "On Atmel M0, M4, 328p or 32u4 it's on GPIO 10"
https://learn.adafruit.com/adafruit-adalogger-featherwing/pinouts

[manic glacier]

tannewt

[adafruit/circuitpython] Pull request opened: #1331 Switch SAMD51 back to \-Os

It messes up neopixel timing otherwise.

Fixes #1326

[tulip sleet]

@slender iron was @idle owl using the debug build?

[slender iron]

no, it was -O0 for all builds

maybe I should pragma that code to always be -Os

[tulip sleet]

🤦

[neon briar]

Thanks @exotic pumice

[tulip sleet]

could put a warning in the Makefile

[slender iron]

https://gcc.gnu.org/onlinedocs/gcc/Function-Specific-Option-Pragmas.html

I like -O0 for debugging

@tulip sleet adding the pragmas now

[tulip sleet]

i'll wait for the travis-rerun

[slender iron]

I confirmed it works even if the top level is -O0

kk

canceled these builds

[manic glacier]

dhalbert

[adafruit/circuitpython] New comment on pull request #1329: Support OSError subclasses and attributes

Unfortunately, since this makes the SAMD21 builds throw different exceptions, this is a pretty user-visible incompatibility. We don't really have room for this in the M0 builds now.

@tannewt notes that each new class adds 60 bytes for the class object, and the most of rest could be qstrs. He has some ideas about shrinking the class objects.

tannewt

[circuitpython:tannewt\-patch\-4] 1 new commit

a3a690d Add pragma to ensure neopixel_write is always -Os - tannewt

[slender iron]

I think the m0 issue is separate. moving onto it now

[tidal kiln]

@neon briar i double checked via the pinouts here:
https://learn.adafruit.com/adafruit-adalogger-featherwing/pinouts#sd-and-spi-pins-2-11
as sajattack mentioned. but i do think the part of the guide you linked is unfortunately confusing. it looks like it might be re-used from the Adalogger Feather main boards (i.e., not the Wing)

[manic glacier]

tannewt

[adafruit/circuitpython] Pull request opened: #1332 Clear Trellis NeoPixels on board reset\.

This makes it easier to change code in cases where the pixels may
cause a brownout.

dhalbert

[adafruit/circuitpython] Pull request closed: #1331 Switch SAMD51 back to \-Os

dhalbert

[adafruit/circuitpython] Issue closed: #1326 Trellis M4 latest build causing issue with NeoPixel code

dhalbert

[circuitpython:master] 3 new commits

ed9db80 Switch SAMD51 back to -Os - tannewt
a3a690d Add pragma to ensure neopixel_write is always -Os - tannewt
c63376c Merge pull request #1331 from adafruit/tannewt-... - dhalbert

dhalbert

[adafruit/circuitpython] New comment on pull request #1332: Clear Trellis NeoPixels on board reset\.

uint8_t zeroes[96] = {0}; works to init to zero also, and the compiler will choose the best way. Don't have to change it if you don't want to.

[digital slate]

got the project built without issue (a first for me, I'm normally chasing compile ghosts)

created a new project

or new board, sourced from the feather m4 project

built with the memory I am using

...and now serial is gone too 😃

devices pop up whenever i'm in the bootloader

but back in the normal CircuitPython mode, it goes dark and I don't see anything

I assume that means something else is wrong

currently on my Linux machine (where I built the project), going to swap back to my windows machine to see if it sees something different on the board.

[idle owl]

Hmm

That's definitely worth trying first.

[tulip sleet]

if you have a crashing main.py or boot.py, that could prevent CIRCUIPTY from appearing

[digital slate]

just loaded the file here back onto my board: https://github.com/adafruit/circuitpython/releases/tag/3.1.1

and it at least comes up with serial

so maybe something wrong with my build?

[tulip sleet]

do you have a SPI flash chip?

this is a homemade board?

[digital slate]

yeah, homemade board with a different chip

that's why I was building

should I be building at a particular tag/release/commit?

[tulip sleet]

what chip?

[digital slate]

https://www.digikey.com/product-detail/en/winbond-electronics/W25Q16JVSSIQ/W25Q16JVSSIQ-ND/6193769

was talking with @slender iron about it above

[tulip sleet]

-IQ and -IM variants have differences. Do we have a def for that chip?

[digital slate]

but even the default feather_m4 wasn't showing serial when I plugged it in

the one I built, that is

[slender iron]

There have been some usb issues so it might be worth trying from before the big usb merge

[digital slate]

should I just chekcout from the release? https://github.com/adafruit/circuitpython/commit/fa1edb2a01f7238ac6fc108142aae35faf7f9b15

that one?

that's the commit tied to the release

[tulip sleet]

yes, try 4.0.0-alpha.2 if you want 4.0. I thought you were using 3.1.1 all the time, but if 4, yes

[digital slate]

nope, not using 4 yet

I mean

[tulip sleet]

3.x branch should be fine

[digital slate]

...am I?

haha

[tulip sleet]

if you're on master, you're on 4

[digital slate]

ah, ok

yeah, then probably need to swap back to make sure

[tulip sleet]

master is pretty unstable right now

[digital slate]

ok, good to know

[manic glacier]

dhalbert

[adafruit/circuitpython] Pull request closed: #1332 Clear Trellis NeoPixels on board reset\.

dhalbert

[circuitpython:master] 2 new commits

4ae4cc1 Clear Trellis NeoPixels on board reset. - tannewt
f190bb2 Merge pull request #1332 from tannewt/clear_tre... - dhalbert

[digital slate]

that'll be a good sanity check against the files I downloaded from the site

(which work a bit more)

and 3.x should build no problem?

or are there other things I need to do other than swap branches

[tulip sleet]

after you checkout 3.x, do a git submodule update --init --recursive

and rebuild mpy-cross if you are freezing anytying

[digital slate]

got it, thanks

(bet that's it)

[manic glacier]

ladyada

[adafruit/circuitpython] New comment on pull request #1332: Clear Trellis NeoPixels on board reset\.

nice!

[digital slate]

woooot

CIRCUITPY shows

as is the case of many woes in my current life, I didn't pay enough attention to the BOM

😃

but this was a useful journey of learning how to build the firmware as well

[idle owl]

@digital slate 🎉

[digital slate]

when the feather boards are shipped from the adafruit factory, they have main.py programmed onto that drive, is that part of the build process as well? Should I be seeing those files or would I need to include them in future build processes?

[slender iron]

They aren’t on there by default. The test code that loads circuitpython loads them onto the flash

[digital slate]

that makes sense

fancy system y'all got going around here 😃

thanks again for troubleshooting help

[manic glacier]

notro

[adafruit/circuitpython] New comment on pull request #1329: Support OSError subclasses and attributes

Version 4

• Drop subclasses

As expected, a lot less flash use:

Base:
58368 bytes free in flash out of 499712 bytes ( 488.0 kb ).

OSError patch: Use 384 bytes
57984 bytes free in flash out of 499712 bytes ( 488.0 kb ).

uerrno patch: Use 328 bytes
57656 bytes free in flash out of 499712 bytes ( 488.0 kb ).

[solar whale]

@slender iron @tulip sleet congratulations on the trellisM4 fix -- just built and ran on mine -- working well now! Nice!

[slender iron]

@solar whale any usb problems?

[solar whale]

not that I have noticed - on the trellism4

[slender iron]

k cool maybe its the m0 that has gremlins still

[idle owl]

Ok so.... long story short the character LCD library never had support for the RGB shield, and was using incorporated driver shims for the other versions.... I redid the entire thing to use MCP230xx, sorted both I2C and SPI for the backpack, and I added code that makes the shield work. I've literally never done anything like this. I'm so crazy excited. (Had to share.)

[solar whale]

I will try more boards tomorrow - or possibly later tonight.

Ive been writing to and running some test scripts on the trellism4 -- not using neopixels -- but no issues with USB

[manic glacier]

notro

[adafruit/circuitpython] New comment on pull request #1323: shared\-module/os: Fix os\.mkdir\('a/b'\)

CPython passes the path straight through to (g)libc (posix_mkdir()):

static PyObject *
posix_mkdir(PyObject *self, PyObject *args, PyObject *kwargs)
{
...
        result = mkdir(path.narrow, mode);
...
}

glibc passes it straight through to Linux (mkdir()):

int
__mkdir (const char *path, mode_...

[manic glacier]

notro

[adafruit/circuitpython] New comment on pull request #1323: shared\-module/os: Fix os\.mkdir\('a/b'\)

@Makdaam Why do you strip off backslashes in #1066?

[slender iron]

@tulip sleet I caught a concurrency issue similar to what you saw. ```
(gdb) p *_usbd_q
$20 = {buffer = 0x20000401 <_usbd_qdef_buf> "", depth = 16, item_size = 12, count = 0, wr_idx = 9,
rd_idx = 8, overwritable = false, mutex = 0x0}

count != wr_idx - rd_idx

and it should

[ruby lake]

@tulip sleet I can now see the usb serial port of the Trellis m4, but the CIRCUITPY: usb drive does not exist.

[manic glacier]

sommersoft

[adafruit/circuitpython] Pull request opened: #1333 Update Core Drivers Page \[3\.x\]

master PR soon to follow.

sommersoft

[adafruit/circuitpython] Pull request opened: #1334 Update Core Drivers Page \[master\]

📑

[tulip sleet]

@ruby lake the visibility of CIRCUITPY shouldn't depend on the drivers, so that is some other problem. are you using the trellis m4 build linked in the trellis guide?

[ruby lake]

checking

[ruby lake]

I updated that using a pi

hm, I thin kthe reset button on this board is broken

it has no tactile feel and the dotstar led glitches a lot

[ruby lake]

@tulip sleet ok, things are better now. I have the drive appearing and serial console

hm, actually it does not see the REPL in Mu

[manic glacier]

sommersoft

[adafruit/circuitpython] New comment on pull request #1333: Update Core Drivers Page \[3\.x\]

Yeah, I had restarted the pca10056 build once before. I haven't sent anything this way in a while, so not sure if 3.x builds are perpetually broken for that specific board. 🤷‍♂️

@dhalbert, any insight?

sommersoft

[adafruit/circuitpython] New comment on pull request #1334: Update Core Drivers Page \[master\]

Argh! Good catch! I shall return my previously received 🌮

[tough flax]

No joy on the RFM69 transmit/receive

😦

One chip is the built-in M0, the other is a wing on an M0 Bluefruit (loaded the Adalogger bin)

Temperature: 22.0C
Frequency: 914.999mhz
Bit rate: 250.0kbit/s
Frequency deviation: 250000.0hz
Sent hello world message!
Waiting for packets...

Built-in output:
Temperature: 32.0C
Frequency: 914.999mhz
Bit rate: 250.0kbit/s
Frequency deviation: 250000.0hz
Sent hello world message!
Waiting for packets...

Temp is off

on the built-in

I'm too tired - will chat w/@tulip sleet about this in the morning

Nevermind! My bad - 😃 All is working

(except the temp is wrong)

[manic glacier]

ATMakersBill

[adafruit/circuitpython] Issue opened: #1335 Order of import statements impacts memory?

The following code succeeds in running:

from adafruit_hid.keyboard import Keyboard
from adafruit_hid.keyboard import Keycode
import adafruit_rfm69
#code to do the work...

while the following fails:

import adafruit_rfm69
from adafruit_hid.keyboard import Keyboard
from adafruit_hid.keyboard import Keycode
#code to do the work...

I get the following error:

Traceback (most recent call last):
  File "main.py", line 10, in <module>
  File "adafruit_hid/keybo...

[manic glacier]

tannewt

[adafruit/circuitpython] Issue closed: #1335 Order of import statements impacts memory?

tannewt

[adafruit/circuitpython] New comment on issue #1335: Order of import statements impacts memory?

It can because the memory gets fragmented differently depending on allocation order and the size of objects. Loading mpy files uses less memory so its recommended to do that for files you aren't editing.

[hoary hill]

If I need to drive a stepper motor directly via DRV8833 breakout board, which circuitpython library to use? I only found examples and libraries for Arduino, not circuitpython...

[stuck elbow]

you may need to write your own

[solar whale]

@slender iron @tulip sleet FYI -- tried current master on pca10059 -- still not working. Will try on some atmel M0 and M4's next

[solar whale]

@tough flax FYI -- The CP build to use for the Feather_Bluefruit_M0 is indeed the Feather_m0_adalogger build, The pin D8 is not available in the feather_M0_basic build and it is needed to se the Chip select for the Bluefruit SPI interface. D8 is avaliable in the adalogger build. Mystery solved...

[solar whale]

@tulip sleet @slender iron No luck with current master on any M0 builds yet -- tried feather_m0_express,feathere_m0_adalogger,, trinket_m0,CPX --- none boot to REPL.

Metro M0 express "looks" like it boots -- green pulsing neopixel but no CIRCUITPY drive ....

Metro M4 express seems to be working fine -- so so far only M4s (MetroM4 and TrellisM4) are working for me.

[manic glacier]

jerryneedell

[adafruit/circuitpython] New comment on issue #1327: latest master does not boot for feather\_m0\_adalogger

FYI - tried commit f190bb2c269b06ab2ce8ee8aa220682ab4f2fc2f on several boards,
the MetroM4 express and TrellisM4 both appear to be working well.
None of the M0 boards tased will boot to present CIRCUITP
tried feather_m0_express,CPX,feather_m0_adalogger,metro_m0_express
Also tried pca10059 -- no luck there either.

[river quest]

one for @ruby lake for later (#comingsoon) ...

[tidal kiln]

@hoary hill check this out: https://circuitpython.readthedocs.io/projects/motor/en/latest/api.html#adafruit-motor-stepper

[tidal kiln]

@tulip sleet another flash erase related thread:
https://forums.adafruit.com/viewtopic.php?f=57&t=143622&p=709201#p709003
note they tried to use the UF2s from here:
https://github.com/adafruit/Adafruit_SPIFlash/tree/master/examples/flash_erase_express
which apparently did not work.
are they different than the ones offered here?
https://learn.adafruit.com/welcome-to-circuitpython/troubleshooting#old-way-for-the-circuit-playground-express-feather-m0-express-and-metro-m0-express-18-25

[manic glacier]

ladyada

[adafruit/circuitpython] New comment on issue #1335: Order of import statements impacts memory?

@kattni this does come up once a while, please add it to our FAQ :)
https://learn.adafruit.com/circuitpython-essentials/circuitpython-expectations#frequently-asked-questions-18-8

[slender iron]

@solar whale thanks for the update. I know what I need to fix for the internal flash variants of nrf but have been focused on the M0s. One thing worth trying on the M0 is leaving it plugged in and hitting the reset button. I believe I'm seeing a clock startup issue from time to time

[solar whale]

I did try that on some. No response. I can try again in a little while.

[slender iron]

k, I'm going to try other m0s today after breakfast. some of them may be glaring issues

I'm seeing some crazy clock stuff on the metro sometimes. I have a code.py with 500hz pwm and it can be in the 600s after. The neopixel doesn't work then either

[idle owl]

Anyone done much with character LCDs? I'm thinking there's a superfluous property in the RGB lib for it but I want to make sure I'm not missing something.

[ruby lake]

@river quest woo

[stuck elbow]

@idle owl the rgb lib doesn't do character lcds?

[idle owl]

@stuck elbow I mean the character_lcd_rgb.py file

For character LCDs with RGB backlights

[stuck elbow]

ah, no idea about that one, sorry

[idle owl]

Ok thanks!

[manic glacier]

Makdaam

[adafruit/circuitpython] New comment on pull request #1323: shared\-module/os: Fix os\.mkdir\('a/b'\)

The code doesn't really strip anything (the path is left intact) it only accepts "\\x00" and "/\x00" as valid endings of a path (before that it was only looking for chars lower than the printables - I don't know why not only \x00).

I added the rule for backslash as well as slash, because it was already used as a separator in other parts of the oofatfs code. No other rationale behind it.

[proven bluff]

If notro is here, sorry for replying so late.

[tulip sleet]

@idle owl I use char lcds but haven't done anything with the rgb ones

[idle owl]

@tulip sleet It's an RGB specific question. I think I'm right but if anyone else had worked with them I wanted an opinion.

@tulip sleet How are you at schematics? Maybe you could at least take a look and see if you agree. I don't have a schematic for the display itself, only it used on a shield, but it shows the pins on the display which is the thing I'm questioning

[tulip sleet]

i am good at schematics, so ga

[idle owl]

https://cdn-learn.adafruit.com/assets/assets/000/036/490/original/lcds___displays_schem.png?1476373463

There is no "backlight" pin on that display. Because the backlights are the red green and blue pins.

Correct?

[tidal kiln]

think of it as 3 separate backlights

[tulip sleet]

right, there are three backlights instead of just one. The anode should be the common +V for all the backlights. You ground the R and/or G and/or B pins to turn them. I have one to verify

[idle owl]

Ok.

[tulip sleet]

well not that one, but I have the raw display

[idle owl]

Because there is a thing setting the the specific backlight pin, and then it uses set_backlight to turn it on or off. I don't think it belongs in this file, because it's meant for the monochrome.

That's what I was thinking.

[tulip sleet]

which pin?

[idle owl]

It defaults to None

So it's not actually setting anything.

But that means the property is in there, doing nothing

and you could add it to your code, and nothing would happen

because I don't think it makes sense for this LCD

[tulip sleet]

for a monochrome backlight, do you give it the pin to ground or the +V pin?

[idle owl]

Ok, monochrome and RGB are two different files. But here is what I used for the monochrome pins and it works: https://cdn-learn.adafruit.com/assets/assets/000/035/681/original/lcds___displays_schem.png?1473701923

Has a backlight pin.

so set_backlight makes sense and also works

[tulip sleet]

would you set_backlight to pin 16?

[idle owl]

let me look

No it's set based on the expander chip and the shift register

I'm using external libraries for both.

[slender iron]

@proven bluff notro is github only

[idle owl]

the MCP230xx and the soon-to-be 74HC595

using the get_pin() property

[tulip sleet]

which library file should I look at?

[proven bluff]

thanks @slender iron

[idle owl]

It's not all pushed to GitHub, hold on

There aren't any issues with the monochrome backlight

My question was regarding the lack of a separate "backlight" pin on the RGB one, separate from red, green and blue.

but let me get as much of this as I can pushed to GitHub

it's still a mess.

[tidal kiln]

@idle owl sounds you're asking about backlight here?
https://github.com/adafruit/Adafruit_CircuitPython_CharLCD/blob/master/adafruit_character_lcd/character_lcd_rgb.py#L132

[idle owl]

yes. and then https://github.com/adafruit/Adafruit_CircuitPython_CharLCD/blob/3775cd7a033b3a9cf2c6806d77807570ffb67794/adafruit_character_lcd/character_lcd_rgb.py#L263 this

[tidal kiln]

not sure, but looks like you just ignore that maybe...

see here:
https://learn.adafruit.com/character-lcds/python-circuitpython#python-and-circuitpython-usage-7-12

you create the RGB instance with:

lcd = adafruit_character_lcd.Character_LCD_RGB(lcd_rs, lcd_en, lcd_d4, lcd_d5, lcd_d6, lcd_d7, lcd_columns, lcd_rows, red, green, blue)

so backlight will just be None

[idle owl]

Yes I know

[tidal kiln]

and then the idea is to use set_color

[idle owl]

Right

but that means that the property set_backlight does nothing.

And you could put it in your code, your code will run, and it will do nothing.

I don't think it needs to be in that file.

[tidal kiln]

from what i see, i agree

they're all generally the same in terms of backlight - one or more LEDs that you can PWM control. RGBs are basically just monochrome with 2 extra LEDs

[gusty kiln]

in my mind, the only reason to have it is probably if you want to be able to drop either of an rgb or monochrome display into code that talks to the interface.

and that's already kind of... you're gonna have to know which class you're instantiating anyway.

[idle owl]

@tulip sleet this is what I was referring to: https://github.com/adafruit/Adafruit_CircuitPython_CharLCD/blob/3775cd7a033b3a9cf2c6806d77807570ffb67794/adafruit_character_lcd/character_lcd_rgb.py#L263
But here's the current state of mine. https://github.com/kattni/Adafruit_CircuitPython_CharLCD/tree/mcp230xx-support/adafruit_character_lcd

[solar whale]

@slender iron pressing reset a few times did get Metro_m0 to boot to CIRCUITPY!

[tidal kiln]

or to allow simple code to treat the RGB as a monochrome version, for simplicity

[idle owl]

But there's no physical pin

so

How would it do anything?

[tidal kiln]

but then the RGB version of set_backlight should operate on the rgb_led

[solar whale]

But no REPL

[slender iron]

@solar whale ok interesting. did the neopixel work each time?

is this with ubuntu?

[solar whale]

yes neopixel pusling green

[idle owl]

which I considered @tidal kiln

but then it's not set_backlight

and it shouldn't be called the same thing in both and do different things

[tidal kiln]

@idle owl this is one of those things where i'd ask the original author if possible

[tulip sleet]

i agree w/ cater and brennen. one thing to worry about is the polarity. The code talks about R and G and B anodes (+), but the shield schematic you posted above has a common anode pin, so the polarity is opposite.

[solar whale]

yes ubuntu 18.04 -- I can see the CIRCUITPY foloer -- looks OK and screen /dev/ttyACM0 connects, but not response from REPL

[idle owl]

@tidal kiln Not possible.

[tidal kiln]

😦

[idle owl]

Already looked into that when I started this.

[tidal kiln]

then you are free to decide what the future holds 🌈

[idle owl]

@tulip sleet the polarity thing is already an issue because set_backlight works backwards at the moment.

[solar whale]

pressing RESET does drop the /dev/ttyACM0 to the metro_m0

[idle owl]

ok. I'm taking it out. or maybe just have it throw an error in case someone uses the same code as from a monochrome

[solar whale]

but it came back to CIRCUITPY

[tulip sleet]

i'd just take it out. It will throw an error because there's no property. That' good enough.

[tidal kiln]

i could argue to leave it in

[idle owl]

The error I'd throw is "This isn't supported for RGB. Instead use set_color"

[tidal kiln]

allow an RGB display to act like a monochrome

[tulip sleet]

well, there's also an optional arg to __init__. Do you want to support that? I'd say if you want to make it act like a monochrome, use the monochrome class, not the RGB class.

you can then specify which pin you want. Unless the polarity issue complicates things also

[idle owl]

no, I would take all of the backlight stuff out.

[slender iron]

@solar whale if you have time can you try it with this version of tinyusb? https://github.com/tannewt/tinyusb/tree/careful_reset

[tidal kiln]

have RGB extend monochrome?

(it doesn't currently)

[idle owl]

I don't get what the use case is here.

This seems like an edge case. That could be done using the current code anyway

[tulip sleet]

would that make the rgb class smaller?

[tidal kiln]

you could run monochrome code with an RGB display, and it would just act like a monochrome display

R=G=B=backlight

[idle owl]

and what color would the backlight be?

[tidal kiln]

shades of white

[idle owl]

That will fail on RPi I think.

[tidal kiln]

or whatever was set with set_color

[solar whale]

@slender iron OK will give it a try -- take me a few minutes to figure out how...

[idle owl]

Can't use more than one LED at the same time

[tulip sleet]

you'd have to make all three pins be the backlight pin, can't do that

[tidal kiln]

i'm thinking overriding backlight in RGB class

[slender iron]

@solar whale you can go into the submodule and add me as a remote

[solar whale]

ah

[tidal kiln]

@idle owl just a guess - but current character_lcd_rgb.py was probably started from character_lcd.py copy pasta style, that's why the set_backlight code is there and the same between the two

[idle owl]

Yes I agree.

But since this is my first major driver work, I wanted to make sure I wasn't missing something before making the change.

[tidal kiln]

i don't think so

[solar whale]

@slender iron it now has ```origin https://github.com/tannewt/tinyusb.git (fetch)
origin https://github.com/tannewt/tinyusb.git (push)

that was git remote -v

[tidal kiln]

are you planning on moving to @property style also?

[idle owl]

yeah.

I picked this up yesterday... I only got it working entirely this morning.

[solar whale]

can I just do git fetch origin careful_reset?

[slender iron]

ah, ya. I think that will work

[solar whale]

seemd to accept it -- building now

[slender iron]

k I'm trying on adalogger too

[tidal kiln]

@idle owl fwiw - it's easier when you start from scratch. dealing with "why is this here?" kind of stuff when inheriting code is an unfortunate side effect. you end up burning time reverse engineering and speculating about things.

[solar whale]

on metro M0 -- I had to reset it twice htne it boots OK and I have REPL

[slender iron]

was the neopixel ok when it didn't work?

[solar whale]

commit did not change -- I htink I have ther right build ```uto-reload is on. Simply save files over USB to run them or enter REPL to disable.

Press any key to enter the REPL. Use CTRL-D to reload.
Adafruit CircuitPython 4.0.0-alpha.2-149-gf190bb2c2 on 2018-11-14; Adafruit Metro M0 Express with samd21g18

[slender iron]

seems like it should be marked dirty

git status should show a change in tinyusb

[solar whale]

checking

[slender iron]

(make sure you checkout the branch too)

[solar whale]

HEAD detached at 299a2f12
nothing to commit, working tree clean
```   ahh - I did not checkout -- trying again

[slender iron]

👍

whats the bossac command?

[solar whale]

just asec

./bossac -e -w -v -R -p ttyACM0 adafruit-circuitpython-feather_m0_adalogger-0.8.0.bin

[slender iron]

thanks!

[solar whale]

hmmm -- no REPL on metro_m0 now -- didi get CIRCUITPY after two resets....

[slender iron]

hrm

[solar whale]

git status ```n branch master
Your branch is up to date with 'origin/master'.

Changes not staged for commit:
(use "git add <file>..." to update what will be committed)
(use "git checkout -- <file>..." to discard changes in working directory)

modified:   ../../lib/tinyusb (new commits)

Untracked files:
(use "git add <file>..." to include in what will be committed)

../nrf/drivers/

no changes added to commit (use "git add" and/or "git commit -a")

[slender iron]

the adalogger worked for me when I loaded it

[solar whale]

OK -- I can try it -- did metro_m0 work for you?

[slender iron]

it was pretty reliable yesterday. when it failed the neopixel did too

[solar whale]

ddi you do debug build -- I am not

[slender iron]

no, let me post my bin

[solar whale]

adalogger not working for me either 😦

[slender iron]

thats for adalogger

[solar whale]

do you have a UF2?

I have the update bootlaoder on mine

[slender iron]

ya

I need to do that

[solar whale]

hmm -- no luck with your build either.

ok after reset -- get CIRCUITPY

and repl

[slender iron]

hrm

so it sounds better but not perfect

[solar whale]

Press any key to enter the REPL. Use CTRL-D to reload.
Adafruit CircuitPython 4.0.0-alpha.2-150-ga96580e92-dirty on 2018-11-14; Adafruit Feather M0 Adalogger with samd21g18
>>> 
>>> 
>>> 

[slender iron]

I think I'll PR these changes so we can see

[solar whale]

code runs!

[tulip sleet]

I will try on a metro m0 as well

[slender iron]

👍

[manic glacier]

iayanpahwa

[adafruit/circuitpython] New star added

[slender iron]

having a code.py with pwm out can help diagnose clock startup issues because a saleae can verify its speed

[solar whale]

my metro_m0 still not happy -- do you want me to try your .uf2?

[slender iron]

let me build one

[tulip sleet]

not working for me on metro m0 with adafruit/circuitpython f190bb2c269 (head) and tannewt/tinyusb at tip of careful-reset 😦

rebuilding again to double-check

[solar whale]

same for me

@tulip sleet try a reset or two

[tulip sleet]

i did, and also plug/unplug

[slender iron]

hrm, is the neopixel working?

[tulip sleet]

yes, pulsing

non-debug build

[slender iron]

so its not a clock issue

[solar whale]

@tulip sleet and I are consistent

[tulip sleet]

on USB2 and USB3 ports

Ubuntu 18.04

i can try windows and mac too

[slender iron]

k let me get closer to a clean branch

[solar whale]

on thirdt reset I got a CIRCUITPY mount, but still not REPL

[slender iron]

@solar whale do you have a beagle?

(usb debugger)

[solar whale]

no -- my dog is part beagle 😉

[slender iron]

😃

[tulip sleet]

do you want another beagle trace from me?

[solar whale]

I get repl on mac

Press any key to enter the REPL. Use CTRL-D to reload.
Adafruit CircuitPython 4.0.0-alpha.2-149-gf190bb2c2-dirty on 2018-11-14; Adafruit Metro M0 Express with samd21g18
>>> 
>>> 
>>> 

[slender iron]

sure I guess

[solar whale]

after a reset

[slender iron]

making a new build for myself now

[solar whale]

@slender iron are you on mac?

[slender iron]

metro

ya, but I also have an arch linux box I test with

[solar whale]

does it work on both? works on my mac but not linux

that was my build -- I can try yours

[slender iron]

ya, I think so

[tulip sleet]

worked on mac first time, but then I unplugged/plugged, and am now getting an irregular green flicker on the neoxpixel on metro m0

and no repl or circuitpy

[slender iron]

that sounds like a clock issue

[tulip sleet]

yah

maybe not clock, maybe excessive interrupts, since it is green

[solar whale]

mine (my build ) was OK after plug.unplug on mac -- still needed a reset

[slender iron]

here is my code.py: ```
import pulseio
import board

p = pulseio.PWMOut(board.D8, frequency=500)
p.duty_cycle = 2 ** 15

while True:
pass```

@chrome perch can you repro with a jlink attached?

oops wrong dan

[tulip sleet]

i'll try, and you want me to monitor D8 with saleae?

[slender iron]

ya, please

[tulip sleet]

you want me to do this on mac or linux?

[slender iron]

either as long as you plug it into what causes the issue

[solar whale]

I can hook it up too -- seeing solig green on mac after reset

[tulip sleet]

the flicker is new, never seen this before, only on mac

[slender iron]

it should be very close to 500 hz

[solar whale]

I see same thing -- solid green on Ubuntu, but no boot

[slender iron]

what linux kernel is it?

[tulip sleet]

Linux salmonx 4.15.0-38-generic #41-Ubuntu SMP Wed Oct 10 10:59:38 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux

[slender iron]

~ $ uname -a
Linux dalinar 4.18.16-arch1-1-ARCH #1 SMP PREEMPT Sat Oct 20 22:06:45 UTC 2018 x86_64 GNU/Linux

[solar whale]

mine is ame as Dan

[slender iron]

k, I may go try it with the rosie box which is running ubuntu

[tulip sleet]

let's see if i can use jlink on linux with it plugged into Mac

[slender iron]

thats what I do 😃

[tulip sleet]

we are running ubuntu 18.04, updated to latest kernel

[slender iron]

k will update my box. its a little behind

[tulip sleet]

i mean latest 18.04 kernel

[solar whale]

my salea is showwing 500hz -- but no USB

on Ubuntu

[slender iron]

k, so not a clock problem

do you have wireshark?

[solar whale]

dmesg shows ```[245044.317406] usb 3-3.4: new full-speed USB device number 111 using xhci_hcd
[245044.440146] usb 3-3.4: New USB device found, idVendor=239a, idProduct=8014
[245044.440152] usb 3-3.4: New USB device strings: Mfr=2, Product=3, SerialNumber=1
[245044.440156] usb 3-3.4: Product: Metro M0 Express
[245044.440158] usb 3-3.4: Manufacturer: Adafruit Industries LLC
[245044.440161] usb 3-3.4: SerialNumber: D5B6D46D8413D405C413E263011241FF
[245044.440937] usb 3-3.4: can't set config #1, error -32

[tulip sleet]

that's the same error I was getting

[solar whale]

I do not have wireshark

[slender iron]

ok, thats good its the same error. I was digging in the code that responds to the set config request

I don't know where its getting lost though

ok, I got it too.

gonna move that box down here

[tulip sleet]

plugged into mac, removed jlink (because it interferes with neopixel), runing the PWM test code. PWM is fine at 500 Hz. LED started regular green, then switched to very bright green after a few seconds. can't reproduce that yet

[slender iron]

hrm

[tulip sleet]

stil not working on win10 either

[slender iron]

@tulip sleet I've reproed the beagle trace you sent me. lots of stalls

[tulip sleet]

a usb state viewer says "Configuration Descriptor Request Failed", which is what I saw in the usb traces

[slender iron]

ya, its weird. it looks like it works the second time

[tulip sleet]

maybe windows and linux are too fast? You could compare the time stamps on the enumeration requests/responses in the good and bad traces and see whether one is significantly ahead of the other?

[solar whale]

I can't get it to work at all on Ubuntu now -- still get 500hz on D8 but no response from USB

arg - unplugged/plugged and CIRCUITPY came right up!

[slender iron]

I've got it reproing with gdb attached

[solar whale]

but no REPL

[slender iron]

I kinda suspect a buffer overrun

the setup request data isn't making it through correctly

[tulip sleet]

but there is something on Mac and Arch Linux that's different: order of requests or timing related

that makes it less likely to fail

[meager fog]

@umbral dagger heya i replied in github but dunno if you didnt see it (you mentioned before not getting github notes) - wanna split Debouncer out?