other than that it just installs a non-specific version of black, so you can skip that step once you follow CircuitPython instructions for uncrustify on mac
#circuitpython-dev
idle owl
ok
onyx hinge
thanks for asking instead of blindly typing your password ๐
idle owl
Thank you for clarifying!
lone axle
Is there a quick way to check (or a cached place to look) for which languages on a given build result in the largest build size?
i.e. which language(s) are most important to test when looking at a PR in a library that is frozen in to some of the smaller builds.
onyx hinge
@lone axle sadly, no.
lone axle
Happy to have confirmation. I'll check a few more
onyx hinge
Here is the notes document for next Mondayโs 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 youโll be attending the meeting - itโs super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and weโll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/1fJ0odbK7m4jWHTnKL1medCeZAGBTZiNIW0np6nOd4hc/edit?usp=sharing
idle owl
Ooof. It reformatted SO many files. I had to git stash and drop after copying my changes out and paste them back in (because I don't know how to do things the right way). Mine was not reformatted though, so it must be correct.
onyx hinge
huhhhhh that's odd
I thought theirs was the same as ours ๐ was it .c files or .py files?
idle owl
Py.
Started looking at git diff and it was going on forever, so I q'd out of it, and blew away the changes.
I guess that means it didn't touch my file. ๐
I can't get to the beginning of it, too much scrollback. But, it ended with the Black format message.
"Both C and Python code are auto-formatted using the tools/codeformat.py script."
I guess it did C first?
I can apparently pass arguments to only do the changed file, but there's no instructions on how to do that.
main furnace
If I connect via the uart, can I use rshell with the broadcom port of CircuitPython?
crimson ferry
how do you move an issue from one library to another? (https://github.com/adafruit/Adafruit_Learning_System_Guides/issues/2115 I think should go to Requests)
manic glacierBOT
idle owl
how do you move an issue from one library to another? (<https://github.com/adafr...
"Transfer Issue" at the bottom of the right hand menu. If you don't have it, I can transfer it.
crimson ferry
I don't seem to have that one
idle owl
I'll transfer it then.
crimson ferry
thank you!
idle owl
thank you!
Done! Thank you for identifying the need for the transfer.
slender iron
If I connect via the uart, can I use rshell with the *broadcom* port of CircuitP...
I don't see why not
crimson ferry
@idle owl follow-up Q: if a Learn Guide is broken due to a library issue, do we want to track the issue in both places? If the library fix is long-term, it may help the guide to switch to a more friendly data source (in the present example)
idle owl
<@!330227457296957440> follow-up Q: if a Learn Guide is broken due to a library ...
Hmm, yeah. File an issue on the Learn repo and mention the Requests issue. That would be a good idea.
crimson ferry
ok, I'll do that
idle owl
Oof. I can't get this I2C power pin to turn off. I thought my initial more-complex code was wrong, but I'm trying the basic code, and setting it True or False has the same results - the STEMMA QT connector still has power, and data is being read from the sensor.
@tulip sleet Even with the pin being turned on by default in CircuitPython (presumably on startup), shouldn't I be able to control it in code.py? Or am I missing something obvious here.
tulip sleet
which board is this?
idle owl
Feather ESP32-S2, rev B.
(I assume it's rev B, there's no marking on the back, it's one of the original boards.)
tulip sleet
loooking...
idle owl
I'm running absolute latest CircuitPython to have Limor's fancy code to start up both rev B and C.
Point was to duplicate it in Python. Which I think I did. But it's not affecting the pin.
tulip sleet
i need to look at the rev B schematic; I have it in eagle
idle owl
Ok
When the duplicated fancy code didn't work, I went back to basics, and that isn't doing anything either.
slender iron
@tulip sleet when do you want to chat?
tulip sleet
<@!329766224093249548> when do you want to chat?
i have something in 45 minutes, could chat for a shorter time
slender iron
I was planning on going for a run
free tomorrow between 11am and 4pm pacific
(or later after my run)
tulip sleet
i might be free by 5pm your time, but that's dinner. tomorrow afternoon my time is good; taking Jesse to a long appt in the morning, back around the time you'll be free. tomw is probably better
we can say 11am your time 2pm my time tomw
slender iron
๐
tulip sleet
i'll put it in my calendar
slender iron
(other folks are welcome to join too. we'll meet here on discord)
idea is to go over stuff before I'm out
blissful pollen
(other folks are welcome to join too. we'll meet here on discord)
My time is kinda hit and miss but if there is anything core related (and overly time sensitive) that could use another set of hands or review while you're gone feel free to reach out. I think by this point I have somewhat of an idea about it
slender iron
Thanks Mark! I don't think there is. Your reviewing helps a bunch though.
tulip sleet
@idle owl could you post your code? I will try barebones with current CPy and also maybe one before Limor's code fix
idle owl
import time
import board
import digitalio
import adafruit_mcp9808
# i2c_power = digitalio.DigitalInOut(board.I2C_POWER)
# i2c_power.switch_to_output()
# time.sleep(0.1)
# rest_level = i2c_power.value
# print(i2c_power.value)
# i2c_power.switch_to_output(value=(not rest_level))
# print(i2c_power.value)
i2c_power = digitalio.DigitalInOut(board.I2C_POWER)
i2c_power.switch_to_output()
i2c_power.value = True
mcp9808 = adafruit_mcp9808.MCP9808(board.STEMMA_I2C())
while True:
temperature_celsius = mcp9808.temperature
temperature_fahrenheit = temperature_celsius * 9 / 5 + 32
print("Temperature: {:.2f} C {:.2f} F ".format(temperature_celsius, temperature_fahrenheit))
time.sleep(2)
Fancy code is commented out.
Was trying to get it to work. When it seemed like it was in terms of making things opposite, but nothing was happening with the pin, I switch to original.
So, no idea if I even did the commented out bit properly, but it looks like what's in CircuitPython.
tulip sleet
I only have the TFT one, but I think it should be the same. I'll check the schematic
idle owl
Um... I'm not certain it is the same. I mean, you should still be able to turn off the pin, but I feel like there was a difference. Could be wrong though.
tulip sleet
VSENSOR part looks the same
idle owl
Ok, then maybe it is
Except the pin is TFT_I2C_POWER on that one, I thought. Combined for the TFT and I2C.
I guess the logic should be the same though.
tulip sleet
yes, it does control both the BME and the TFT, but the TFT is just another connection to VSENSOR
idle owl
Ah ok
tulip sleet
ok, hmm, on the TFT, the I2C pullups are powered by VSENSOR, but on the plain rev B, they are always powered. I think at least some sensors may end up being powered by the pullups by accident
that would explain it
idle owl
oof.
Ok.....
Well then. I guess I'll ask Limor what she wants to do. Means you can't force low power with a sensor plugged in? Or maybe it stops talking to it in deep sleep anyway..... I don't know.
I have to go fix veggies for dinner. I'll pick this up tomorrow. Thanks for looking into it.
tulip sleet
that is part of the redesign. Yes, ask Limor
there are at least two low-power problems with the <rev C plain Feather
idle owl
Hmm... ok
Heading out. Talk to you tomorrow for specifics. Or read back here to find them if you post here.
crimson ferry
@idle owl if all else fails, we could have a rev B pins.c and a rev C pins.c, and check hasattr for _INVERTED
manic glacierBOT
misty garnet
is asyncio pretty lightweight or does it add significant overhead?
shut dew
Looks like this release needs its PyPi release action triggered https://github.com/adafruit/Adafruit_CircuitPython_Wiznet5k/releases/tag/1.12.3 vs https://pypi.org/project/adafruit-circuitpython-wiznet5k/
solar whale
Does anyone know how we can tell what boards support the BLE workflow? VIa code.circuitpython.org I can connect to a feather bluefruit sense but I am also trying to use a Maker Diary nrf52840 USB Dongle and I don't seem to be able to get it to show up. It does not show the blue status LED on boot and may well not support the BLE workflow but I was curious how to look at the build to determine what boards do support it.
lone axle
Looks like this release needs its PyPi release action triggered <https://github....
It looks like none of the release actions triggered when the release was made. That is odd, not really sure what could cause it. I think I recall mention of actions having trouble the other day, maybe this release was made during downtime and it got "lost to the ether" and never triggered once everything was back to working.
lone axle
@idle owl do you think making a new release with no code changes for this: https://github.com/adafruit/Adafruit_CircuitPython_Wiznet5k/releases/tag/1.12.3 is the right thing to do in order to trigger the release actions to run? I think pypi has issues if you try to release the same code again to it. But in this case it never made it to pypi or the release assets page.
idle owl
<@!330227457296957440> do you think making a new release with no code changes fo...
It's worth a try. Worse cast we delete everything, edit something, and do another PR/release. I see it's not on the release actions page at all. So this might work.
lone axle
ty, I'll give it a shot.
lone axle
Looks like this release needs its PyPi release action triggered <https://github....
I made a new release and it does seem to have triggered the actions successfully this time. Latest version should be deployed on pypi now. Thanks for mentioning this issue.
idle owl
@solar whale Merged! https://github.com/micropython/micropython/pull/8406
solar whale
<@!343570430042308609> Merged! https://github.com/micropython/micropython/pull/8...
Yay!!! Congratulations!
idle owl
Thanks! ๐
solar whale
You are now a Micropython core developer!!
idle owl
Hah! I don't know about all that ๐
idle owl
You are now a Micropython core developer!!
Limor said basically the same thing within a minute of your message ๐
She used "contributor" which I guess is vaguely more accurate.
@solar whale Anyway, if you test it, let me know. I'm going to add a page to the Feather ESP32 V2 guide on using MicroPython. Which I will probably ask for your help with.
solar whale
<@!343570430042308609> Anyway, if you test it, let me know. I'm going to add a p...
I'll try the "nightly build" hopefully tomorrow morning
idle owl
Excellent! Thank you!
idle owl
I did not venture into that yet. I was going to wait for a build on their end as well.
solar whale
I thought so too -- but do I have to tap reset as it is starting? I tried, but I have not yet been able to see it in code.circuitpython.org
No big deal -- I'll keep trying
solar whale
no that blinks blue.. just a red led
slender iron
I think it should have different blink rates
that's just a power led though I think
the board def doesn't have a status led defined
solar whale
I may just have to play with the timing. Good to know it "should" work.
solar whale
the board def doesn't have a status led defined
right -- so I guess I just guess at the timing
slender iron
the page says it has an rgb led
solar whale
it does
solar whale
I'll add it to my "todo" list! Be fun to try. but that should not prevent it from working woth BLE should it?
slender iron
no, I don't think it should
but it'll make it easier to understand what is going on
solar whale
OK -- sorry -- I have to run. I'll take a look at it later today or tomorrow to see I fI can make any progress -- as well as look into adding the LED
slender iron
np. thank you!
crimson ferry
Does a library in / override a frozen library?
tulip sleet
Does a library in `/` override a frozen library?
In the usual sys.path, it comes first, so yes, it should
crimson ferry
thanks, came up in a library issue
lone axle
That is super useful! I didn't realize there was a way to make it not use the frozen one.
tulip sleet
@slender iron I am available now, but see you are working with Limor.
happy to wait for that to finish
slender iron
I'm in the circuitpython voice channel
tulip sleet
@onyx hinge we are going over issue lists before Scott leaves, if you have time, in the public CircuitPython voice channel
if not we'll just do it the two of us
idle owl
That is super useful! I didn't realize there was a way to make it not use the fr...
Oh yeah, we needed a simple way to iterate on things. This was a decision we made ages ago for that exact purpose.
idle owl
So apparently at once point, wifi had reset() but now it does not? I'm not sure what you would use in place of that. Since I have no idea what it previously did.
It was suggested that I use this loop in my current example. I have it almost working properly, but it fails on this with no attribute found, and the documentation backs that up.
crimson ferry
@idle owl closest thing is:py wifi.radio.enabled = False wifi.radio.enabled = True wifi.radio.connect(secrets['ssid'], secrets['password']) # _added later_
idle owl
Hmm ok.
crimson ferry
less intrusive would be to stop the station, then start it and re-connect
idle owl
I'm kind of fuzzy on what I was trying to do anyway. Chances are I will be handing this off to someone else. I'm 20 minutes shy of the timebox I was given for it, and moved on to something else because I was kind of going in circles.
crimson ferry
(there's is no explicit wifi disconnect or reset (native esp32-s2), but supervior.reset would be the next escalation)
tidal kiln
@lone axle did this end up being the fix for the docs?
https://github.com/adafruit/Adafruit_CircuitPython_PortalBase/pull/66
i lost track of that whole discussion
lone axle
<@!382939733107408897> did this end up being the fix for the docs?
<https://gith...
I think that PR didn't specifically fix it. Because it was still empty after that go merged.
tidal kiln
yah...seeing the second PR now too...
lone axle
My leading theory is that this one fixed it: https://github.com/adafruit/Adafruit_CircuitPython_PortalBase/pull/67 it it was weird. It got fixed on the live page as soon as I made this PR rather than waitng for it to get merged, which definitely seems odd to me. So possibly not correct.
tidal kiln
ha! ok. so maybe some mystery still to it.
oh well. the docs are up now at least.
thanks for looking into that!
lone axle
You're welcome
misty garnet
is RS485 (rs485_dir keyword) support planned for the SAMD or RP2040 boards? There should also be a note in the docs indicating the it currently is only working for the ESP32 boards
manic glacierBOT
weary hill
Hello all. I just released Piku v0.2.0. It will now automatically install a package dependencies, you can also now install specific versions of packages, it creates a lock file like poetry/npm etc, and there is an update command to update all packages to latest.
split ocean
say @lone axle i'm trying to install vectorio_helpers w circup, but it's throwing a warning WARNING: vectorio_helpers is not a known CircuitPython library.
ah looks like they just aren't in bundle yet. I've installed them by hand, and the line works well, about to try outlined rectangle.
crimson ferry
Is there any CircuitPython mdns example code? I'm having issues, but not sure which side of the keyboard they're on ๐
lone axle
ah looks like they just aren't in bundle yet. I've installed them by hand, and t...
Yep, you ended up on the right conclusion. Not in the bundle yet so circup doesn't see them. I'll add them to the CircuitPython Org bundle this week though and after that circup should be able to find.
I may still end up tweaking the way that OutlinedRectangle works as well to try to make it as memory efficient as possible.
slender iron
Is there any CircuitPython `mdns` example code? I'm having issues, but not sure ...
# SPDX-FileCopyrightText: 2022 Scott Shawcroft for Adafruit Industries
# SPDX-License-Identifier: MIT
"""
"""
import mdns
import microcontroller
import socketpool
import wifi
from adafruit_httpserver import HTTPServer, HTTPResponse
print("connecting")
wifi.radio.connect("skylab", "freewifi")
print("Connected")
pool = socketpool.SocketPool(wifi.radio)
server = HTTPServer(pool)
m = mdns.Server(wifi.radio)
print(f"Listening on http://{m.hostname}.local:80")
m.advertise_service(service_type="_circuitpython", protocol="_tcp", port=80)
for service in m.find(service_type="_circuitpython", protocol="_tcp", timeout=5):
print(service.service_type, service.protocol, service.hostname, service.instance_name, service.port)
@server.route("/temperature")
def base(request): # pylint: disable=unused-argument
"""Return the current temperature"""
# pylint: disable=no-member
return HTTPResponse(body=f"{str(10)}")
# Never returns
server.serve_forever(str(wifi.radio.ipv4_address))
Thanks! I got finding working, though it safemodes if there are no services of a given type.
slender iron
Please file an issue. I may not have time to fix but it'll be good to track.
slender iron
split ocean
@lone axle does this look like it should work? open_box = outlined_rectangle.OutlinedRectangle(pixel_shader=palette4, width=10, height=6, x=40,y=300, outline_color_index=0, outline_thickness=4)
i'm getting an error File "outlined_rectangle.py", line 48, in __init__ TypeError: extra keyword arguments given
lone axle
<@!382939733107408897> does this look like it should work? `open_box = outlined_...
Ah, that is because of the color_index that it's passing to Rectangle super class constructor. That is brand new. Just merged into core within the past few days. But if you grab the latest build from S3 for your device I think it should be in there.
lone axle
Indeed! it's a great way of drawing basic shapes without Bitmap overhead.
blissful pollen
you can even say vectorio really connects the dots...
split ocean
i'm playing around with it making this little circus game thing, based on some bouncing ball code from @devout jolt
lone axle
Oooooh fun! that is coming out fantastic
split ocean
Thanks! Here it is in motion
blissful pollen
that's cool, and now i have that classic circus music stuck in my head
split ocean
haha so sorry. I just learned recently that that was written as a military march, and boy would the composer be disappointed at how it now means ๐ช ๐คก to us
blissful pollen
ha that's a cool random fact. I mean sorta has things choreographed to it still
split ocean
@lone axle I've updated to S3 latest (Feather RP2040 w TFT FeatherWing for hardware) and the error has shifted! File "outlined_rectangle.py", line 61, in __init__ TypeError: 'module' object is not callable
lone axle
<@!382939733107408897> I've updated to S3 latest (Feather RP2040 w TFT FeatherWi...
That one is a stumper, not sure. Ill take a look in the morning. Possibly i changed something in the module on my device that didnt get checked in
split ocean
right on, not urgent! Thanks much.
pallid echo
Lady Ada mentioned that in a revision of one of the Feathers she fixed the deep sleep current draw. Was that something related to these internal pull-downs and external pull-ups If it is, my intuition says we could fix this in software for the older revision of the board. Sorry if I remembered incorrectly.
I bumped into this because I wanted to see whether it was feasible to go into deep sleep in the asyncio library on the MagTag when all tasks are busy. Maybe by pickling the co-routines into RTC memory. I really like the async API.
manic glacierBOT
solar whale
@idle owl I tried the 20220323 nightly build for MP on the feather esp32 v2 and I no longer get the invalid Pin error with scl = Pin(20)) but I am not able to access my I2C device ```MPY: soft reboot
network config: ('10.0.0.161', '255.255.255.0', '10.0.0.1', '75.75.75.75')
MicroPython v1.18 on 2022-03-23; ESP32 module (spiram) with ESP32
Type "help()" for more information.
import pos
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "pos.py", line 31, in <module>
File "lsm6dsox.py", line 95, in init
File "lsm6dsox.py", line 152, in __read_reg
OSError: [Errno 116] ETIMEDOUT
I tried a simple test -- connecting Pin 22 (SDA) to Pin 13 (LED) then ```from machine import Pin
p = Pin(22,Pin.OUT) --- the RED LED comes on
p.value(False) --- the RED LED goes off
then move to Pin 20 (SCL)
p = Pin(20,Pin.OUT) --- no response on RED LED
p.value(False) --- no response on RED LED
p.value(True) --- no response on RED LED
also tried Pin 14 -- works as expected -- LED responds
also reading back pin 20 after setting it to True or False. It always returns 0. Other pins return the expected value>>> p = Pin(20,Pin.OUT)
p.value()
0
p.value(True)
p.value()
0
p = Pin(22,Pin.OUT)
p.value()
0
p.value(True)
p.value()
1
wow is github having trouble yet again today? I'm getting a lot of unicorn pages accessing actions results. githubstatus is all green though ๐
sterile crag
I am looking for a program that runs on MacOS to log CircuitPython output. Thonny has a scrollback buffer of only about 1000 lines I can select with <command>A. Mu has a larger buffer, but no SelectAll shortcut. On Windows I use putty, but I can't/don't want to install putty on MacOS. I tried rshell, but it only runs with MicroPython (It needs ubinascii.unhexlify.) Likewise screen does not seem to work with CircuitPython.
TLDR: I need output logging on MacOS and CircuitPython.
jaunty juniper
I am looking for a program that runs on MacOS to log CircuitPython output. Thonn...
I like tio, there are others like picocom or python -m serial.tools.miniterm
screen works with Circuitpython, but I don't even know how to get the scrollback if possible, though there should be a way to output to a file automatically
tulip sleet
Lady Ada mentioned that in a revision of one of the Feathers she fixed the deep ...
VSENSOR did not control the on-board pullups. Now it does
There was a different issue with the QT Py ESP32-S2, which has also been fixed.
jaunty juniper
tio -log logfile /dev/cu.usbmodemBLABLABLA for example will connect and log all to a file
or just something like cat /dev/cu.usbmodem146443133 > output.log will log all to a file
onyx hinge
@jaunty juniper ooh I didn't know about the -log option of tio, I'll remember it!
sterile crag
@jaunty juniper Thanks. tio -loutput.txt /dev/tty.usbmodem487F307DBD4D1 works. Didn't think of the cat solution.
tio is nice because I can use the REPL.
lone axle
@split ocean I tested out that outlined rect code you posted here last night and I am seeing the box drawn on the display. I put the code I ended with here: https://gist.github.com/FoamyGuy/296abd0d2800e1790b8c76276838ab27 I tried to copy the constructor from your message and just fill in the minimum other code to run on RP2040 + 3.5 FeatherWing. I did need latest from S3, but after that I'm seeing it get drawn.
If your still seeing the module object not callable when you get back to it, maybe try this bare bones version to see if you still get that error. And if possible you can send the code you have to me and I can try it on my end to see if I can track down the specific cause. Looking back at the error message and the lines of code referenced I'm not seeing any immediate issues jumping out, but could still be something that is somewhat environment dependent that I am missing.
manic glacierBOT
split ocean
@lone axle I tried that code to the same error -- maybe this has to do with me running outlined_rectangle.py and line.py as their own libraries and not from the vectorio_helpers library
lone axle
<@!382939733107408897> I tried that code to the same error -- maybe this has to ...
that could be. Inside of outlined_rectangle.py the import it uses for Line assumes that it will be inside of vectorio_helpers You could change the import at the top of outlined_rectangle if you wanted though and it should be able to work that way I think.
split ocean
i had already made that change to the Line import in outlined_rectangle
lone axle
line 29 in outlined_rectangle.py is currently:
from vectorio_helpers.line import Line
but something like this should work I think if you moved it out on it's own:
from line import Line
Hmmm, then I am less sure. I would have expected it to work like that but maybe my understanding of how the imports work is not quite accurate.
split ocean
ahah! i had done that wrong. I had made it import Line oops ๐
@lone axle ok now it works! thanks.
lone axle
Ah. I'm surprised it wasn't a module not found error. So there is definitely still a gap in my understanding. Glad that got it going for you though ๐
split ocean
Yay! Say, do you think things like rotation and scaling could be possible one day in vectorio?
lone axle
Yay! Say, do you think things like rotation and scaling could be possible one da...
Funny you should mention it. Your carnival looking example last night actually gave me the idea of a RotatedPolygon helper.
I think it will involve some geometry to do the rotation function on the given points in the polygon. So it'll take a bit of math to do but I think we can make a fairly easy to use helper class.
stuck elbow
rotozoom
lone axle
ah that does exist as well. in bitmaptools, though I don't think it can be used on vectorio shapes.
split ocean
Yes, i wasn't about to solve those points in my head as i built the flags so i made them in Rhino, rotated them, and used Grasshopper to export the vertex coordinates to bring into CircuitPython
jaunty juniper
Ah. I'm surprised it wasn't a module not found error. So there is definitely sti...
it looks like import is not case sensitive (is fat case sensitive ?)
stuck elbow
fat is not case sensitive
lone axle
I will make an attempt at a helper class for it in the next few days and add it to the repo if it ends up working how I am imaging.
jaunty juniper
so import Line imports line.py
split ocean
oh @lone axle one other question related to the outlined_rectangle -- is there a way to have the rectangle fill be transparent and the stroke opaque?
related question -- is there a reason the outlined_rectangle seems to use a three color palette with first color for outline and third color for fill? I'm not sure what the second color is there for.
lone axle
related question -- is there a reason the outlined_rectangle seems to use a thre...
No specific reason. I was partially testing out the new color_index option from the core.
lone axle
oh <@!382939733107408897> one other question related to the outlined_rectangle -...
I don't think it can right now. But I'm not certain. Possibly somehow with Palette.make_transparent() That is definitely some functionality I'd love to have though I'll look into ways to make it happen.
split ocean
right on, thanks
idle owl
<@!330227457296957440> I tried the 20220323 nightly build for MP on the feather...
Oof. I have no idea how to begin troubleshooting that. I see you posted to the issue. Might have wanted to make a new issue, but I don't know how Damien likes things.
solar whale
Oof. I have no idea how to begin troubleshooting that. I see you posted to the i...
Yeah, I was not sure how to handle that. I canโt do anything else on this today. I have no idea what is wrong. If you can reproduce it, that would be good. It would be great if I am just doing something wrong.
idle owl
Yeah, I was not sure how to handle that. I canโt do anything else on this today....
It's on my list. Will be the first time using MicroPython... so a lot of variables where things could go wrong on my end unrelated to this. ๐
I will give it a try though.
solar whale
Good luck!
idle owl
Thanks!
slender iron
unicorn when getting build output is frustrating
idle owl
I noticed the Python Discord folks had a custom badge for their Discord chat on their various GitHub repos. I had Bruce design one for us. We can host it in the Bundle repo, and link to it that way in all of the Adafruit CircuitPython libraries. It would replace the "chat" badge. This won't show the total number of folks online like the current chat badge does, but I think that's fine. If anyone else has opinions on this, please let me know.
idle owl
@solar whale I made the LED blink using the MicroPython REPL!
Also, I guess I'm getting similar results: >>> i2c = I2C(1, scl=Pin(20), sda=Pin(22), freq=400000) E (1051500) i2c: /home/micropython/esp-idf-v4.2/components/driver/i2c.c:767 (i2c_set_pin):scl gpio number error
Or an error, anyway.
๐ Not sure what to do with this. I wanted to try to do a simple scan, but it won't even initiate properly.
Wondering if there's some place where "valid" pins for I2C etc are called out explicitly, and this didn't get added there.
Also I have no idea how to upload a... main.py I guess it is. I got to the REPL and made the board do a thing, but I'm unclear on using MicroPython past that ๐
jaunty juniper
ampy --port /dev/tty.something put "main.py"
stuck elbow
close your repl for it
jaunty juniper
yep, you can also change the name with put main_test_file.py main.py for example, so you can test different files without renaming them locally
idle owl
Does MP have a serial console like CP?
It won't break out of the REPL for me at the moment, so I wasn't sure.
jaunty juniper
it works like CP, except no autoreload
jaunty juniper
also, you can use Thonny or Mu, enable the "file" panel, it will automatically switch to and from the REPL as needed, though uploading files is a little awkward with those
idle owl
I downloaded the generic ESP32 build, but the REPL is reporting a specific board name. I have to head out for a bit. I'll look further into this when I'm back.
I verified my download... so I'm not sure what to make of this.
slender iron
@idle owl do we have a standard pin name for battery voltage?
idle owl
<@!330227457296957440> do we have a standard pin name for battery voltage?
Um... let me think for a minute.
I'm not sure we do for battery voltage, but let me check.
idle owl
<@!330227457296957440> do we have a standard pin name for battery voltage?
This is on the Feather M4. { MP_ROM_QSTR(MP_QSTR_VOLTAGE_MONITOR), MP_ROM_PTR(&pin_PB01) }, { MP_ROM_QSTR(MP_QSTR_BATTERY), MP_ROM_PTR(&pin_PB01) },
What Dan said.
idle owl
YW.
Hokay, so, MicroPython. REPL says MicroPython v1.18 on 2022-03-23; LILYGO TTGO LoRa32 with ESP32 I downloaded the generic ESP32 version, and it's reporting that board.
Wondering if that's why the pin assignment is failing.
But not sure how to get the right version if the website isn't providing what I need.
I mean, yeah, build it, but I don't know how to do that.
mental nexus
@slender iron thanks for the pointer to FrameBuffer. FrameBuffer experts: Are there any good examples showing to how a framebuffer gets hooked up to a physical display (eg how a display is initialized and connects the display board to a framebuffer) and then how displayio and framebuffer are setup to talk together.
slender iron
I know when I use it on broadcom, it handles the framebuffer to display bits
it is also used for matrix displays
and protomatter handles the framebuffer
solar whale
@idle owl I think you want the esp32 spiram build for the v2
idle owl
<@!330227457296957440> I think you want the esp32 spiram build for the v2
I saw that in the issue you filed, and downloaded it and flashed it just now.
Type "help()" for more information.
>>> from machine import Pin, I2C
>>> i2c = I2C(1, scl=Pin(20), sda=Pin(22), freq=400000)
E (69290) i2c: /home/micropython/esp-idf-v4.2/components/driver/i2c.c:767 (i2c_set_pin):scl gpio number error```Still getting the same error.
solar whale
Use I2C 0
solar whale
Yes
idle owl
Same error.
idle owl
I did that earlier, and it did work.
I can do it again though, hold on
Worked this time too.
solar whale
I donโt know why it works for me.
idle owl
I'm simply trying to do the I2C scan to see if the device shows up. Can't even get that far. ๐
solar whale
Just canโt use it
idle owl
Ah.
I copied that line from the "Getting Started with ESP32" page. Am I doing it wrong?
solar whale
Iโll be home in about 15 min. Iโll check mine again.
idle owl
Ok, thank you!
onyx hinge
@idle owl this does work in arduino?
idle owl
<@!330227457296957440> this does work in arduino?
Which? I2C? or...?
solar whale
I have not tried it
onyx hinge
haha / argh
idle owl
Scan at least worked in Arduino.
onyx hinge
Author: Alberto Garciฬa Hierro <alberto@frsky-rc.com>
Date: Wed Apr 21 22:01:14 2021 +0100
Enable IO20 on ESP32
Some newer ESP32 variants (like ESP32-PICO-V3 and ESP32-PICO-MINI-02)
do implement this pin and it can be used as a normal GPIO.
Fixes #6016
Fixes #6837
Closes https://github.com/espressif/esp-idf/pull/6918
diff --git a/components/soc/esp32/gpio_periph.c b/components/soc/esp32/gpio_periph.c
index a413203741..d993cb42ac 100644
--- a/components/soc/esp32/gpio_periph.c
+++ b/components/soc/esp32/gpio_periph.c
@@ -35,7 +35,7 @@ const uint32_t GPIO_PIN_MUX_REG[SOC_GPIO_PIN_COUNT] = {
IO_MUX_GPIO17_REG,
IO_MUX_GPIO18_REG,
IO_MUX_GPIO19_REG,
- 0,
+ IO_MUX_GPIO20_REG, // This corresponding pin is only available on ESP32-PICO-V3 chip package
IO_MUX_GPIO21_REG,
IO_MUX_GPIO22_REG,
IO_MUX_GPIO23_REG,
diff --git a/components/soc/esp32/include/soc/soc_caps.h b/components/soc/esp32/include/soc/soc_caps.h
index ae88415cb4..03597fe989 100644
--- a/components/soc/esp32/include/soc/soc_caps.h
+++ b/components/soc/esp32/include/soc/soc_caps.h
@@ -116,8 +116,8 @@
// set pullup/down/capability via RTC register. On ESP32-S2, Digital IOs have their own registers to
// control pullup/down/capability, independent with RTC registers.
-// 0~39 except from 20, 24, 28~31 are valid
-#define SOC_GPIO_VALID_GPIO_MASK (0xFFFFFFFFFFULL & ~(0ULL | BIT20 | BIT24 | BIT28 | BIT29 | BIT30 | BIT31))
+// 0~39 except from 24, 28~31 are valid
+#define SOC_GPIO_VALID_GPIO_MASK (0xFFFFFFFFFFULL & ~(0ULL | BIT24 | BIT28 | BIT29 | BIT30 | BIT31))
// GPIO >= 34 are input only
#define SOC_GPIO_VALID_OUTPUT_GPIO_MASK (SOC_GPIO_VALID_GPIO_MASK & ~(0ULL | BIT34 | BIT35 | BIT36 | BIT37 | BIT38 | BIT39))
diff --git a/docs/en/api-reference/peripherals/gpio.rst b/docs/en/api-reference/peripherals/gpio.rst
index 9f92f8d5be..93a5e05523 100644
--- a/docs/en/api-reference/peripherals/gpio.rst
+++ b/docs/en/api-reference/peripherals/gpio.rst
@@ -118,6 +118,11 @@ Overview
-
-
-
+
+ * - GPIO20
+ -
+ -
+ - This pin is only available on ESP32-PICO-V3 chip package
* - GPIO21
-
```a change is needed in esp-idf for it to work, such as this one. It's in esp-idf v4.4 tag but not in older ones. which version is actually used to build micropython?@onyx hinge wild guess. Did you erase the flash before loading MP?
onyx hinge
I don't even have one of these boards ๐
solar whale
I meant @idle owl
idle owl
<@!426200547024961539> wild guess. Did you erase the flash before loading MP?
I did, both times.
solar whale
Shucks
idle owl
```commit 6deaefde69c3f068d39b76d628b05b21fd142b79
Author: Alberto Garciฬa Hierr...
The scl gpio error says esp-idf-v4.2
idle owl
Oof.
onyx hinge
# build the boards, based on the IDF version
for board_json in $(find boards/ -name board.json | sort); do
mcu=$(cat $board_json | python3 -c "import json,sys; print(json.load(sys.stdin).get('mcu', 'unknown'))")
if idf.py --version | grep -q v4.2; then
if [ $mcu = esp32 ]; then
# build standard esp32-based boards with IDF v4.2
if echo $board_json | grep -q GENERIC; then
# traditionally, GENERIC and GENERIC_SPIRAM boards used manifest_release.py
MICROPY_AUTOBUILD_MAKE="$MICROPY_AUTOBUILD_MAKE FROZEN_MANIFEST=$(pwd)/boards/manifest_release.py" build_board $board_json $fw_tag $dest_dir bin elf map
else
build_board $board_json $fw_tag $dest_dir bin elf map
fi
fi
else
if [ $mcu != esp32 ]; then
# build esp32-s2/s3/c3 based boards with IDF v4.4+
build_board $board_json $fw_tag $dest_dir bin elf map uf2
fi
fi
done
oof that paste, but yes they build esp32 with idf 4.2 and the newer chips with idf 4.4. so this won't work without the right idf
idle owl
Hmph. Ok.
solar whale
but why does it work differntly for me -- at my desk -- wiil have an example in a minute
Type "help()" for more information.
>>>
>>> from machine import Pin,I2C
>>> i2c = I2C(0,scl=Pin(20),sda=Pin(22))
>>>```is this the same build artifact that kattni used?
solar whale
but Pin 20 in non functional for me -- Is that related to the IDF?
idle owl
Looks like it....
solar whale
should be -- I will upload the file I loaded
idle owl
Type "help()" for more information.
>>> from machine import Pin, I2C
>>> i2c = I2C(1, scl=Pin(20), sda=Pin(22), freq=400000)
E (69290) i2c: /home/micropython/esp-idf-v4.2/components/driver/i2c.c:767 (i2c_set_pin):scl gpio number error```er
I2C 0 gives the same error
onyx hinge
It's not identifying itself closely, since that's not exactly "v1.18"
idle owl
that was my first try.
Right but the build day is the same, and they only do nightlies, afaict.
solar whale
I can use Pin 20 all I want, but it always reads 0 --- even if I just set it as an output and write a 1 to it
idle owl
Identical hash to what I downloaded.
idle owl
Yes to both.
solar whale
says V2 in big letter
idle owl
Yep ๐
solar whale
I'm baffled
onyx hinge
F? weird, how did you get my report card from 9th grade health class?
solar whale
both as to why you get differnt results and why Pin 20 does not work for me....
idle owl
Agreed.
solar whale
@onyx hinge what is the issue with the IDF version?
onyx hinge
@solar whale they had to add BIT20 to SOC_GPIO_VALID_GPIO_MASK so that it became possible to use pin GPIO20
onyx hinge
idf 4.2 does the check a bit differently but yeah
it has const uint32_t GPIO_PIN_MUX_REG[GPIO_PIN_COUNT] = { and in the middle: ```c
IO_MUX_GPIO18_REG,
IO_MUX_GPIO19_REG,
0,
IO_MUX_GPIO21_REG,
IO_MUX_GPIO22_REG,
I don't know why kattni and jerryn would have different wrong results but I'm pretty sure it's not going to work with today's idf 4.2.
Kattni, you got thrown in at the deep end ๐
idle owl
When am I not?
Is there a possibility for a weird bug? No? TOO BAD. KATTNI WILL FIND ONE.
onyx hinge
where do micropython nightly builds come from?
solar whale
So I guess it is in MP's court to update to 4.4
idle owl
where do micropython nightly builds come from?
Their setup, I believe.
They build and upload to the downloads section of the website.
solar whale
idle owl
I only just now figured out the snek in the MicroPython "M" logo. ๐คฆ๐ปโโ๏ธ
solar whale
I'm not sure what else to do -- except wait for a comment on the issue -- hopefully it will be seen even on a closed issue.
onyx hinge
I literally can't find the script that checks out esp-idf version 4.2 within the micropython github. I can find evidence of 4.0.2 and 4.4. ```tools/ci.sh: ci_esp32_setup_helper v4.0.2
tools/ci.sh: ci_esp32_setup_helper v4.4
I would look at how long it took Damien to respond at the beginning of the issue, and then post a new issue after that amount of time if there is no response. Refer to the closed issue in the new one to link them.
solar whale
but @idle owl maybe you should add your experiences - since they are different. -- feel free to open a new issue and reference the old one if you prefer.
idle owl
but <@!330227457296957440> maybe you should add your experiences - since they ar...
Hah. We're on the same thought process.
I'll open a new one so it's not you reposting the same thing.
onyx hinge
further action is required and the existing issue is closed, so there should be a fresh issue. my feeling.
idle owl
Yeah agreed. Jerry, you can post to the new issue if you like once I file it.
idle owl
It's kind of a new problem though.
Pin 20 is exposed, it's not working for I2C is the issue.
solar whale
can you jsut do p20 = Pin(20,Pin.OUT)
then ```p20.value()
p20.value(True)
p20.value()
that is, is it just I2C
idle owl
E (2562560) gpio: gpio_set_direction(263): GPIO number error```MicroPython v1.18 on 2022-03-23; ESP32 module (spiram) with ESP32
Type "help()" for more information.
>>>
>>> from machine import Pin
>>> p20 = Pin(20,Pin.OUT)
>>> p20.value()
0
>>> p20.value(True)
>>> p20.value()
1
>>>
I thought earlier you said the value was always 0
onyx hinge
earlier when you got 0, was an i2c device connected?
it's remotely possible that a freshly reset device could be have differently too
solar whale
bah ````>>>
MPY: soft reboot
network config: ('10.0.0.161', '255.255.255.0', '10.0.0.1', '75.75.75.75')
MicroPython v1.18 on 2022-03-23; ESP32 module (spiram) with ESP32
Type "help()" for more information.
from machine import Pin
p20 = Pin(20,Pin.OUT)
p20.value()
0
p20.value(True)
p20.value()
1
p20.value(False)
p20.value()
0
p20.value(True)
p20.value()
0
is it just changing randomly? sum(p20.value for _ in range(1000))
solar whale
sum(p20.value() for _ in range(1000))
0
onyx hinge
okay, so it's not randomness
solar whale
as you said may have been a reset condition
oh wait -- I hat it connected to Pin 13
with it disconnected
106
>>> ```I clicked the reset button, and tried to make p20 again, and am getting the same error.
onyx hinge
wait what, connecting something to pin 13 makes pin 20 behave differently?!
I'm going home.
solar whale
106
>>> sum(p20.value() for _ in range(1000))
92
>>> sum(p20.value() for _ in range(1000))
86
>>> sum(p20.value() for _ in range(1000))
93
>>> ```I had pin 20 connected to pin 13 so it could blink the led
idle owl
So I'm not sure what to put in this issue ๐
solar whale
I would just report the errors you see. I'll add my contradicting errors ...
onyx hinge
I am not having a good time imagining how jerryn's getting these results
idle owl
Ok.
onyx hinge
all of kattni's results make sense to me, idf doesn't want to support pin 20
I've tried to build micropython for ESP32 but using IDF 4.4. No idea if this works, and no hardware to test on.
idle owl
Ooh on it.
solar whale
is taht for the SPIRAM
onyx hinge
this is esp32 build-GENERIC/micropython.bin and the ref is v1.18-244-g66fe3d5cb
I don't know if it's spiram or not because frankly I don't understand how esp-idf builds. but spiram is not related to the issue, so a non-spiram build should also load(?)
solar whale
teh V2 is the SPIRAM version
onyx hinge
OK, well, it's GENERIC not GENERIC_SPIRAM
solar whale
Tah tis a different build in the MP site.
idle owl
Trying it anyway, I guess.
solar whale
idle owl
It is sad...
solar whale
I have to go -- sorry -- will check on how you made out later, but I can't do any more on this untill tomorrow morning.
๐ฆ
idle owl
Connecting to serial is displaying this repeatedly, no breaking out of it: rst:0x10 (RTCWDT_RTC_RESET),boot:0x13 (SPI_FAST_FLASH_BOOT) configsip: 271414342, SPIWP:0xee clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00 mode:DIO, clock div:2 load:0x3f400020,len:197668 ets Jul 29 2019 12:21:46
idle owl
I have to go -- sorry -- will check on how you made out later, but I can't do an...
No worries! Thanks for your help!
solar whale
You too!
onyx hinge
OK, boo. 4.4 + GENERIC_SPIRAM doesn't fit in flash. ```/home/jepler/.espressif/tools/xtensa-esp32-elf/esp-2021r2-8.4.0/xtensa-esp32-elf/bin/../lib/gcc/xtensa-esp32-elf/8.4.0/../../../../xtensa-esp32-elf/bin/ld: micropython.elf section .iram0.text' will not fit in region iram0_0_seg'
/home/jepler/.espressif/tools/xtensa-esp32-elf/esp-2021r2-8.4.0/xtensa-esp32-elf/bin/../lib/gcc/xtensa-esp32-elf/8.4.0/../../../../xtensa-esp32-elf/bin/ld: IRAM0 segment data does not fit.
/home/jepler/.espressif/tools/xtensa-esp32-elf/esp-2021r2-8.4.0/xtensa-esp32-elf/bin/../lib/gcc/xtensa-esp32-elf/8.4.0/../../../../xtensa-esp32-elf/bin/ld: region `iram0_0_seg' overflowed by 440 bytes
collect2: error: ld returned 1 exit status
ninja: build stopped: subcommand failed.
Hmm.
onyx hinge
I hope you can get yours esptool'd back to something that doesn't bootloop ๐
idle owl
On it.
solar whale
Thanks for trying @onyx hinge Good evening !
idle owl
I hope you can get yours esptool'd back to something that doesn't bootloop ๐
No worries. It's back.
onyx hinge
It was clearly foolhardy to submit & accept that PR without testing (aside from building) but that's water under a bridge now.
it's so weird to not have CI artifacts from github but they have their own ways I guess
idle owl
Yeah...
Hmm......
>>> i2c = machine.I2C(scl=machine.Pin(21), sda=machine.Pin(22))
Warning: I2C(-1, ...) is deprecated, use SoftI2C(...) instead
>>> i2c = machine.SoftI2C(scl=machine.Pin(21), sda=machine.Pin(22))
>>> i2c.scan()
[]
>>>```The power LED isn't illuminated on the I2C breakout, so it's not getting power from the STEMMA QT connector.
Looking to see which thing is the power pin
onyx hinge
21 & 22 now, not 20 & 22?
oh and yes there's the power pin to activate, micropython won't do that for you
idle owl
Oooooohhh... I did it!
hold on
scanning again
I pasted that... no wonder it worked
I didn't notice the pin number difference.
No, it's failing.
I got power to the breakout, but I2C is failing again.
E (437660) gpio: gpio_set_direction(263): GPIO number error
E (437660) gpio: gpio_set_level(215): GPIO output gpio_num error
>>> i2c = machine.I2C(scl=machine.Pin(20), sda=machine.Pin(22))
Warning: I2C(-1, ...) is deprecated, use SoftI2C(...) instead
E (449630) gpio: gpio_set_direction(263): GPIO number error
E (449630) gpio: gpio_set_level(215): GPIO output gpio_num error```Need the 0 before pin
idle owl
E (528380) i2c: /home/micropython/esp-idf-v4.2/components/driver/i2c.c:767 (i2c_set_pin):scl gpio number error```Itโs backโฆ.
idle owl
So I didn't actually get anywhere. ๐ฆ
Though I apparently learned enough to make a power pin work without help.
onyx hinge
that's still with their nightly, since mine didn't work at all?
idle owl
Yes
solar whale
Progress is not always forward.
idle owl
Here's what I have for the issue:
I submitted the fix to #8393 to expose pin 20 on ESP32. On the Adafruit Feather ESP32 V2, SCL is pin 20. When trying to initalise I2C, it is not accepting pin 20 as a viable option for SCL.
>>> from machine import Pin, I2C
>>> i2c = I2C(0, scl=Pin(20), sda=Pin(22))
E (2540280) i2c: /home/micropython/esp-idf-v4.2/components/driver/i2c.c:767 (i2c_set_pin):scl gpio number error
I2C set up works with pin 14 as SCL.
If I try to simply init pin 20 as an output, it also fails.
>>> from machine import Pin
>>> p20 = Pin(20, Pin.OUT)
E (19500) gpio: gpio_set_direction(263): GPIO number error
It has been suggested that perhaps the issue is resolved in the ESP IDF v4.4, but an attempt to build MicroPython with that version resulted in a failure to fit in flash. So, I have not tested this possibility.
idle owl
Feels like I'm out of my wheelhouse here, so it seems awkward.
onyx hinge
you gotta power past the awkwards, that is all
idle owl
Does that seem acceptable?
onyx hinge
you can say "@onyx hinge suggested that ..." it's ok
idle owl
Ok, thank you
onyx hinge
otherwise no change to suggest
idle owl
@solar whale When you're around, https://github.com/micropython/micropython/issues/8440 is filed.
solar whale
<@!343570430042308609> When you're around, https://github.com/micropython/microp...
Thanks. Iโll repeat my tests and add them in the morning.
idle owl
Thanks. Iโll repeat my tests and add them in the morning.
Thank you!
OK, I need to go dinner before our meeting.
Be back in a bit.
onyx hinge
supervisor.set_next_code_file(__file__, reload_on_error=True) hum this treats KeyboardInterrupt as an 'error' so it tricked me for a bit how to get out of it.
onyx hinge
safe mode via reset button would be one way
ornate breach
Amazing how a misplaced NOT operand can mess up a whole utility
mental nexus
and protomatter handles the framebuffer
I looked through the framebuffer code, and it looks like it creates a memory buffer and displayio (somehow) can write directly into it. If I can expose the pointer to the ESP32-S3 buffer, this might work. However, in the ESP-IDF code, after it writes to the lcd framebuffer in PSRAM, and it calls to flush the cache. Will the framebuffer code flush the cache? https://github.com/espressif/esp-idf/blob/56659cdf36790e182451cd892ea349b29a27d333/components/esp_lcd/src/esp_lcd_rgb_panel.c#L334-L337
mental nexus
Ok, Iโll look at that. Also, I canโt decipher how displayio โknowsโ itโs a framebuffer and writes directly to memory.
slender iron
I think that's FramebufferDisplay's job
the whole group/tilegrid stuff writes into memory with a given format
mental nexus
This has a lot of tentacles, so it will take me a bit to grasp all of them.
slender iron
ya, it's not simple
mental nexus
I need to get out my pinboard and string again.
mental nexus
I think that's FramebufferDisplay's job
Thanks, thatโs a great help! I didnโt realize frameBufferDisplay had its own updating code. I mistakenly assumed it relied on displayio. Thanks for the pointers!
slender iron
it's a peer of Display I believe
inland mural
Hi! I was wondering if anyone can help me with my problem. I have a project for school, and I'm trying to operate 16 individual servos on circuit python using a 16-channel PCA9685 connected to an Arduino nano. Right now, I'm just trying to have one servo on channel 0 move, so that I can test them individually for functionality purposes. However, I'm getting this error on circuit python and I'm not sure if it's a hardware or software problem. Any advice? Attached is my simple code and then the error ๐ซ .
jaunty juniper
Hi! I was wondering if anyone can help me with my problem. I have a project for ...
that's a question for #help-with-circuitpython let's go there
slender iron
circup folks may be interested in pipkin from the creator of thonny: https://forums.adafruit.com/viewtopic.php?f=60&t=189621
lone axle
@split ocean I made some progress on a RotatedPolygon helper class:
Animating very quickly has some clipping artifacts. But for static or relatively slow moving things it will simplify things I think. Perfect for things like the triangle flags across the top of your latest demo
It's pushed into vectorio_helpers with a new minimal example showing it's usage: https://github.com/circuitpython/VectorIO_Helpers/blob/main/examples/rotated_polygon_simpletest.py very early draft though, no docstrings yet.
split ocean
@lone axle nice!
lone axle
Ooh, the previous video accentuates the clipping because it's moving by 10 degrees each frame. Moving by only 1 degree it's much smoother:
split ocean
do you base pivot point on COG?
lone axle
Not sure I know what COG mean
split ocean
sorry, center of gravity, or some other geometric center
(COG was a sort of misnomer that stuck in cg animation)
lone axle
Ah. I think that it rotates around the origin point (0, 0) in the plane with the points it's initialized with.
split ocean
picking an optional pivot point would be highly rad
lone axle
So if you use some negative and some positive starting points it will rotate about the center. I think mine is not quite the real center because I eyeballed the points to use to get it centered-ish.
split ocean
thanks I'll play around with it
lone axle
I'll need to play with it some more to get a feel for how it works. I think it should be possible to make the pivot point more arbitrarily selectable by transforming the points before feeding them to the rotation math.
split ocean
right on, this is exciting
weary hill
@proven garnet I merged your pull request and released a new version with the initialize command. I also added an directory option to the create command and some tests to make sure creating/init in an existing directory didn't copy over existing files (hopefully).
lone axle
@split ocean bedtime soon for me. But I just pushed a tweak to the RotatedPolygon for different colors, and a new example that draws these pretty rainbowy kaleidoscope-ish designs https://github.com/circuitpython/VectorIO_Helpers/blob/main/examples/rotated_polygon_rainbow_circle.py
split ocean
ooh pretty! thanks @lone axle great stuff
solar whale
<@!343570430042308609> When you're around, https://github.com/micropython/microp...
I added my notes -- still no idea why we get different results....
onyx hinge
Does anyone know what board the "cp32-m4" is? It's in the duplicate pid/vid script, but was never documented on circuitpython.org. https://circuitpython.org/board/cp32-m4/
huh added by bryan / siddacious
manic glacierBOT
orchid basinBOT
idle owl
I added my notes -- still no idea why we get different results....
It is odd, for sure.
manic glacierBOT
trim elm
@lone axle Hey can I rename the example/ directory in https://github.com/adafruit/Adafruit_CircuitPython_PortalBase to examples/ ? I'm trying to clean up adabot a bit and it wants the examples directory to be called examples, even if there's just one example
lone axle
<@!382939733107408897> Hey can I rename the example/ directory in https://github...
yes, definitely. It was meant to match all of the other normal libraries, I must have overlooked the plural name.
trim elm
Cool
slender iron
@onyx hinge esp32-m4 was a board that was geared to fitting a SAMD51 in an ESP32 footprint. never was produced
idle owl
Quick show of L or D, who here uses light mode appearance on GitHub, and who uses dark mode appearance?
Pretty close to equal. Interesting, though I was basically expecting as much.
A fix for the reason I asked just popped into my head so before I bother asking for opinions on a thing, I'm going to try my fix.
misty garnet
it doesn't look like there is any 'flush' equivalent for busio.UART? Is this something that's not supported by the hardware?
idle owl
So I resolved the issue with the white border not working on light mode by making it grey or Blinka-purple. But now I'm not sure which one looks best. I also included the original version with no border, which looked fine on light mode, but optical illusion made it look smaller than the other badges on dark mode. Thoughts? I'll upload them separately so folks can use reactions to cast their "votes".
No border:
Grey border:
Blinka-purple border:
slender iron
it doesn't look like there is any 'flush' equivalent for busio.UART? Is this som...
I believe busio.UART writes synchronously. in other words, it flushes immediately
it should all be transmitted when the write call returns
(it might be once the byte is copied to the peripheral)
misty garnet
I had to add a 1ms sleep after the write before I toggled the drive enable line on a RS485 transceiver otherwise I would get bit corruption on the last byte, so it seems like it didn't fully push out that last byte before the line was de-asserted
slender iron
what port?
slender iron
the implementation is here: https://github.com/adafruit/circuitpython/blob/main/ports/atmel-samd/common-hal/busio/UART.c#L373
it's not unheard of to return a little too early
I'm not sure what the SERCOM exposes that we could wait on
misty garnet
looks like the write is non-blocking, but it waits until all bytes are transmitted
does async_status.txcnt get reset to 0 before each transmission?
slender iron
I have no idea. I haven't looked into it
misty garnet
cause it just checks if the transmission count is >= the length of the data before returning
tulip sleet
cause it just checks if the transmission count is >= the length of the data befo...
tx_cnt is incremented when the byte is written to the DATA register for the SERCOM peripheral. So it's not necessarily completely sent. Perhaps we should check usart_async_is_tx_empty() as well in the ASF4 driver API. (that checks an interrupt bit; I'm not sure it reflects that the byte has been completely sent. would need to check the datasheet).
misty garnet
@tulip sleet where is tx_cnt zeroed, before each write?
tulip sleet
circuitpython/ports/atmel-samd/asf4/samd51/hal/src/hal_uasrt_async.c (from ASF4)
it is copied from tx_por
tx_por is zeroed in uasrt_async_write(). also see uasrt_process_byte_sent(), which increments tx_por.
misty garnet
got it, thx
slender iron
@tulip sleet what is the httpserver code based on?
tulip sleet
looking for that...
I think it was this, but I redid it a lot: https://nippycodes.com/coding/circuitpython-esp32-s2-http-server/
slender iron
๐
tulip sleet
It might be more interesting to use https://github.com/deckerego/ampule or https://github.com/Neradoc/circuitpython-native-wsgiserver
i didn't know about either of those beforehand
slender iron
I looked at them
I've run up against the fixed route paths
I was thinking I'd do /fs/<file path>
tulip sleet
i added routing only at the last minute
slender iron
it bugs me that routing isn't just an if/elif
tulip sleet
that sounds fine; one or both of those has file serving in their examples/ directory
slender iron
but no frameworks I can find actually do that
tulip sleet
I agree, I thought that was weird; it's WSGI or nothing
slender iron
I was looking at ASGI too but it uses dictionaries everywhere
tulip sleet
i looked at flask and some other stuff, but they were all too complicated or depended on urllib or whatever
slender iron
pfalcon did picoweb
https://github.com/jczic/MicroWebSrv2 has 428 stars
I couldn't find a common C http server either
ESP-IDF has a custom one
tulip sleet
is too big
you could take the routing arg stuff from one of the tiny WSGI servers and add it on to mine. I have no attachment to what I did; it's just that my first prio was serving a big file, and I had to make that work
slender iron
I think my gut wants to undo the routing stuff
and have an iterator that returns a request
tulip sleet
you can take it out; it's not that functional; it was just an idea I had at the last minute
slender iron
I generally don't like function calls that never return because they hide stuff
tulip sleet
for people who want routing, we can have a WSGI server
slender iron
I mean with other CPython web code. it should be easy to sit on top of CPython sockets
tulip sleet
if you make an iterator instead of serve_forever(), what will each iteration return?
slender iron
a request object
tulip sleet
after it has been served?
tulip sleet
or will you need to call a respond function
jaunty juniper
note that wsgiserver does not call a function that does not return, it polls, you specify an application when setting it up, it's that app that does the routing if any (using adafruit_wsgi in the example)
tulip sleet
so it could serve a file if there is no matching route
it was a bit tricky to get the root of the file serving path right
slender iron
I don't understand the "application" pattern either
tulip sleet
or, like in the example, file-serving is in the app
jaunty juniper
what they call app is an object instance that you can call with the request data
blissful pollen
I'm reading all this and think I really have to read up on wsgi. Been too long since I really looked at stuff like that
tulip sleet
https://peps.python.org/pep-3333/ is the defn of WSGI
jaunty juniper
technically it can also just de a function
tulip sleet
with rationale
jaunty juniper
def do_the_thing(request, start_response):
# do the thing
wsgiServer = server.WSGIServer(80, application=do_the_thing)
wsgiServer.start()
while True:
wsgiServer.update_poll()
Scott, Neradoc would steer you right on this, more experienced than us with this
jaunty juniper
I'm not that experienced though, but the core code is really that simple, it calls the "application", which is a function (or a callable object), whenever there's a request in the poll
slender iron
my struggle with these apis is that they put a layer of code between the code.py code and the users code in a function
which is why I lean towards a loop over requests
I totally get that this is weird for how it is done in normal cpython
jaunty juniper
yeah it's a callback that is automatically triggered by the poll
which is my preference personnally
slender iron
right. I'm not a fan of hidden event loops
jaunty juniper
basically like the neotrellis library versus keypad
jaunty juniper
but your idea can be made into the other if that's what the user prefers too, with a library
slender iron
ya, and they can coexist
slender iron
I like this approach: https://kamilwu.medium.com/a-guide-to-python-asyncio-part-2-a30b9ce918ca
I like that it gives you a connection level function
viscid pine
@slender iron do you have any objections to how i wrote the control flow for installing the uart driver? If needed i can try make it less... like that
slender iron
I thought about it but no since the install params are different
one thing I thought about was if you should just always do the event
viscid pine
it seems more straightforward for sure
and its not like esp platform would care about one extra task doing nothing
slender iron
you just need to ensure you only care about ctrl-c on the debug uart then
but waking the task is ok
misty garnet
is there a list of supported flash chips?
manic glacierBOT
manic glacierBOT
manic glacierBOT
proven garnet
If anyone with the keys to RTD has a chance, adafruit_motor's documentation failed to build
I couldn't find anything small to submit a PR for to get it to retrigger ๐
tulip sleet
If anyone with the keys to RTD has a chance, `adafruit_motor`'s documentation fa...
I am triggering another build
proven garnet
Thanks!
tulip sleet
ok, it worked!
proven garnet
Fixed, thanks again!
proven garnet
The only other related question I have is what's up with https://github.com/adafruit/Adafruit_CircuitPython_Pixelbuf
Does that have an RTD project made for it?
Because it's status is unknown and it brings you to a broken link
tulip sleet
it is missing the webhook; I'll add it
I triggered the motor build directly in rtd
This is confusing: we have https://github.com/adafruit/Adafruit_CircuitPython_Pixelbuf and https://github.com/adafruit/Adafruit_CircuitPython_Pypixelbuf
proven garnet
Yeah, was it that pypixelbuf got changed to pixelbuf?
I can't find usage of pypixelbuf anywhere
I noticed that while I was typing some time ago and meant to follow up haha
tulip sleet
we renamed it to adafruit_pixelbuf in the C implementation. We need to archive one of these, I think. @idle owl ^^
anyway, I will add the other one
proven garnet
Sweet, thanks!
tulip sleet
@idle owl we have pixelbuf and pypixelbuf in the bundle. Can we drop one now because we aren't building <7.x bundles?
I think we would drop pypixelbuf and archive https://github.com/adafruit/Adafruit_CircuitPython_Pypixelbuf, or at least add some warning to its README
proven garnet
That worked, closed the pertinent issue!
jaunty juniper
@tulip sleet it's:
MP_DEFINE_CONST_FUN_OBJ_0(time_not_implemented_obj, time_not_implemented);
(that is swapped in the module globals table based on defines)
tulip sleet
<@!329766224093249548> it's:
```C
MP_DEFINE_CONST_FUN_OBJ_0(time_not_implemented...
hmm, so maybe we should make it a FUN_OBJ_VAR (and change the signature)
jaunty juniper
testing it on QTPY, that seems to take 24 more bytes, maybe not worth it
tulip sleet
or FUN_OBJ_VAR_BETWEEN, with 0 and 1
Does it mess up Thonny?
Right now you don't see the NotImplementedError, which is kinda bad. You see th arg error
It's worth 24 bytes, I think, to make it throw the right thing
jaunty juniper
in Thonny the error is just ignored, says "could not determine epoch, assume 1970" (I don't know why they need it)
onyx hinge
@tulip sleet ^ I'm hoping I did something wrong in my testing setup but it seems the RTC can "overflow" well before the count register actually wraps around. does that make any sense to you at all?
tulip sleet
you mean the overflow interrupt gets set early?
onyx hinge
yes, it seems that way
tulip sleet
i am just looking at the code
onyx hinge
RTC_Handler overflowed with overflow_ticks=0xfffffffd
I saved off the value of the count register inside the interrupt and then print it later outside the interrupt
tulip sleet
i was hoping if we synchronize the read then it should be sync'd up. I didn't read anything in the datasheet that it might be early
so my/your hope/assumption is unfounded
i did some forum searching etc about this and didn't see any cautions. We can open a ticket with Microchip about it too
seems very bad
onyx hinge
adding the 'sync' in the interrupt doesn't markedly change the outcome. ```while ((RTC->MODE0.SYNCBUSY.reg & (RTC_MODE0_SYNCBUSY_COUNTSYNC | RTC_MODE0_SYNCBUSY_COUNT)) != 0) {}
overflow_flag = true;
overflow_ticks = RTC->MODE0.COUNT.reg;
i have to be afk for a bit but will check back with you about this, not sure how to get around it
onyx hinge
just not use the overflow interrupt
onyx hinge
though that's not working for me either, is there a 2nd bug?
tulip sleet
though that's not working for me either, is there a 2nd bug?
you mean you are just checking for wraparound?
onyx hinge
yeah basically.```c
static uint32_t _get_count(uint64_t *overflow_count) {
#ifdef SAM_D5X_E5X
while ((RTC->MODE0.SYNCBUSY.reg & (RTC_MODE0_SYNCBUSY_COUNTSYNC | RTC_MODE0_SYNCBUSY_COUNT)) != 0) {
}
#endif
// SAMD21 does continuous sync so we don't need to wait here.
uint32_t count = RTC->MODE0.COUNT.reg;
if(count < rtc_old_count) {
// Our RTC is 32 bits and we're clocking it at 16.384khz which is 16 (2 ** 4) subticks per
// tick.
overflowed_ticks += (1L << (32 - 4));
}
rtc_old_count = count;
if (overflow_count != NULL) {
*overflow_count = overflowed_ticks;
}
return count;
}
if we avoid caring about exactly when the OVF interrupt fires, we can just look for a wrap in _get_count. Missing one isn't important, unless you don't call a time-fetching function for over 1.5 days.
tulip sleet
are you testing this on samd21 or 51?
onyx hinge
samd51
it's true, we should test on both
hah OK I think I figured out the other bug, it was in my rtc-warping code
tulip sleet
backspacing
onyx hinge
you can write RTC->MODE0.COUNT.reg = 0xffffff00; but it doesn't mean that the value changes immediately
tulip sleet
so maybe the interrupt is not early after all
onyx hinge
OK I'll let this run for a few thousand overflows but I think it's "right" now.
tulip sleet
oh, nvm, you are reading it back as already warped
westfw (who is active on the forums, etc.) noticed some synchronization delays on SAMD10: https://community.atmel.com/forum/samd10-rtc
when the clock speed is slow
i'm grasping at straws
onyx hinge
so does that equation 5รPGCLK + 2รPAPB < D < 6รPGCLK + 3รPAPB mean that reading the RTC COUNT could be up to 6 RTC ticks behind the interrupt? I was seeing up to 3..
onyx hinge
the notification didn't seem to make it to discord but I did open a PR https://github.com/adafruit/circuitpython/pull/6202 I'll test on samd21 after lunch
also no workflows were started for that PR ๐ข
tulip sleet
so does that equation _5รPGCLK + 2รPAPB < D < 6รPGCLK + 3รPAPB_ mean that readin...
That is for SAMD10, not sure what is in the SAMD51 datasheet.
onyx hinge
Same
tulip sleet
so in any case, it can be behind. I think your solution is find, though it does have the need to fetch the time often enough
idle owl
<@!330227457296957440> we have pixelbuf and pypixelbuf in the bundle. Can we dro...
Yes, one can be archived. Whichever one isn't named the same as the built-in version. That was the plan all along, but we couldn't do it immediately.
tulip sleet
it is the pypixelbuf one, substituting for _pixelbuf. ok, sounds good. I will make some PR's.
idle owl
We should literally archive the pypixelbuf repo.
Once we're done making sure it's not referenced.
tulip sleet
yes, I will add a message to it. Or do you want to do it or assign someone?
we just need a PR for the bundle
idle owl
I can archive it, there's a process to it.
tulip sleet
i will look in the learn repo
idle owl
Which I could explain, but could also simply do it in the same amount of time.
tulip sleet
Welcome_to_CircuitPython/code.py
13:* adafruit_pypixelbuf.mpy
Hah ok.
idle owl
```
Welcome_to_CircuitPython/code.py
13:* adafruit_pypixelbuf.mpy
```
Let me know when we're ready to archive. I guess we'll have to manually remove from the bundle? I've never archived a repo that was in the bundle.
tulip sleet
i started to do that, but then discovered a bunch of libraries still refer to it, so multiple more PR's:
halbert@tuna:~/repos/Adafruit_CircuitPython_Bundle$ ag pypixelbuf
libraries/helpers/pixelbuf/README.rst
93:This example tests that pypixelbuf works.
97: class TestBuf(adafruit_pypixelbuf.PixelBuf):
libraries/drivers/neopixel_spi/setup.py
39: "adafruit-circuitpython-pypixelbuf",
libraries/drivers/neopixel_spi/README.rst
25:* `Pypixelbuf <https://github.com/adafruit/Adafruit_CircuitPython_Pypixelbuf>`_
libraries/drivers/neopixel_spi/neopixel_spi.py
34: import adafruit_pypixelbuf as _pixelbuf
39: import adafruit_pypixelbuf as _pixelbuf
libraries/drivers/neopixel_spi/requirements.txt
7:adafruit-circuitpython-pypixelbuf
libraries/drivers/dotstar/adafruit_dotstar.py
25: import adafruit_pypixelbuf as adafruit_pixelbuf
libraries/drivers/is31fl3741/adafruit_is31fl3741/is31fl3741_pixelbuf.py
25: import adafruit_pypixelbuf as adafruit_pixelbuf
libraries/drivers/neopixel/neopixel.py
27: import adafruit_pypixelbuf as adafruit_pixelbuf
circuitpython_library_list.md
280:* [Adafruit CircuitPython Pypixelbuf](https://github.com/adafruit/Adafruit_CircuitPython_Pypixelbuf.git) ([PyPi](https://pypi.org/project/adafruit-circuitpython-pypixelbuf))
docs/drivers.rst
513: PyPixelbuf <https://circuitpython.readthedocs.io/projects/pypixelbuf/en/latest/>
dusk mauve
Did my first esp32s3 build/flash today with an esp32s3-box and the latest main branch. the build/flash worked great! display works, USB-CDC serial console is operating.
Only oddity is that the USB filesystem is corrupted in a weird way I haven't seen before. the image is what it looks like on my Win11 system but the behavior is similar on an ubuntu system as well.
Is this a known issue (or a familiar pattern) or is it worth submitting a bug?
tulip sleet
did you erase flash before flashing the initial stuff?
you can do import storage; storage.erase_fileystem() and see if it fixes it up
in the repl
dusk mauve
did you erase flash before flashing the initial stuff?
knew I forgot something ๐ going to do a full erase reflash
tulip sleet
let us know if that does not work
idle owl
```
halbert@tuna:~/repos/Adafruit_CircuitPython_Bundle$ ag pypixelbuf
libraries/...
Ah yeah. Ok. I'll be around ๐
dusk mauve
you can do `import storage; storage.erase_fileystem()` and see if it fixes it up
Thanks for adding a new tool to my 'wtf is this' debugging toolbox! worked great.
tulip sleet
it's also good when the filesystem gets corrupted due to resetting too soon, etc.
dusk mauve
Thankfully I haven't run into that yet but super good to know.
dusk mauve
build/install from main on two other s3 boards i have on hand (devkitc-1-n8r2 and devkitm-1-n8) went great! lots of good work in the last few months on the s3 port. ๐
lone axle
Is this the builtin font that comes from terminalio.FONT in code? https://github.com/adafruit/circuitpython/tree/main/tools/fonts
stuck elbow
depends on the board, I think this one is used by default, but you can override it in the board definition
lone axle
Thank you, I had recalled seeing on your devices it's a custom one and I found where that was set in the mk files.
idle owl
One final call for which Adafruit Discord badge you prefer! Add a reaction to which one you like better, if you're interested.
Grey border:
Blinka-purple border:
Really only matters on dark mode - they look basically the same on light mode.
tulip sleet
@tidal kiln - neopixel_spi is still at 0.9.3; how do you feel about it being 1.0.0? (I have to make release to change pixelbuf stuff.)
made executive decision to do so, since it's been around for years
idle owl
Version numbers are free ๐
@solar whale So.... MicroPython question. Is there any point in using the nightly build on the ESP32 V2 Feather since pin 20 doesn't work anyway? Or is there still some improvement in the nightly build?
I'm writing a guide page for MP on the V2, and wanted to know if I should be pointing folks to the nightly, or the stable.
tidal kiln
@tulip sleet np. what @idle owl said ๐
solar whale
<@!343570430042308609> So.... MicroPython question. Is there any point in using ...
None that I am aware of.
idle owl
Ok thanks!
tulip sleet
@onyx hinge should rtc_old_count be reset anywhere? It's fine, right because a hard reset will zero it, and the ticks counter is not reset on a soft reload.
The CI job was never scheduled, not sure how to poke it.
onyx hinge
@tulip sleet I rely on it being zero'd at startup. If there was ever a case where we changed the rtc COUNT register we would want to do something about rtc_old_count there, but there is no such place.
onyx hinge
I force-pushed the branch and CI has picked it up now
tulip sleet
excellent, thanks
onyx hinge
I guess you could fix the missed overflow problem (the hypothetical "if you somehow go 1.5 days without calling a time function") by putting the overflow interrupt back and just having its action be to call _get_count and discard the result.
idle owl
Blergh. Pylint issue I've never run into before, and after reading up on the PEP that led to it, I still have no idea how to fix it. I get what the problem is, but I don't even see why Pylint is failing on the line it's failing on, and I definitely don't see how to resolve it. Implicit string concatenation found in list (implicit-str-concat) is the failure. On line 16 in this file. Where is it even concatenating those into a string? I see it's an item in the for, but it's not used in the prints, so... I'm lost.
@jaunty juniper Your dot is green, but it doesn't mean you're around. If you are around... Do you understand this Pylint failure I'm getting and/or have any suggestions to resolve it?
winter mortar
You're missing a comma between 'WPA-PSK' and 'WPA2-PSK4'
idle owl
Really? The code runs. Weird. Trying that.
winter mortar
that's why you're getting the "implicit concatenation" warning
idle owl
๐ Your code has been rated at 10.00/10 (previous run: 9.58/10, +0.42)
jaunty juniper
yeah:
['Open', 'WEP', 'WPA-PSK' 'WPA2-PSK4', 'WPA/WPA2-PSK'] becomes
['Open', 'WEP', 'WPA-PSKWPA2-PSK4', 'WPA/WPA2-PSK']
idle owl
Oof.
Why didn't the code barf when run on the board?
Or if it doesn't need that particular authmode, does it ignore it?
(I didn't write this, I've been tasked with getting it on GitHub.)
jaunty juniper
well actually pylint could also complain about a variable defined but not used since authmodes is not used
so, that it is wrong is irrelevant :p
jaunty juniper
it's not in the code you posted
idle owl
for (ssid, bssid, channel, RSSI, authmode, hidden) in station.scan():
oh.
that's singular.
Wow.
I have no idea why either of them are there, then. Or how the code runs happily. But I don't know MicroPython at all, really, so ...
jaunty juniper
maybe it was originally supposed to match authmode and what is in authmodes, like authmode in authmodes or authmodes[authmode] if authmode happens to be an index...
like some more generic code that was reused
idle owl
That would make a lot of sense. Limor loves reusing other code.
Hmm. I'm wondering what to do with this then. I mean, it works as-is, but that's not great.
I guess start by asking Limor where it came from. It's not in the quickstart for the board, or the docs for the network module.
jaunty juniper
oh yeah authmode is an int
jaunty juniper
that authmodes[authmode] gives you the human readable version
idle owl
Oh ok
jaunty juniper
which explains why it was there
idle owl
except it's not used properly here?
jaunty juniper
it's not used in this code
idle owl
like there's a middle piece missing?
Ok
Trying to figure out how to present this to Limor.
I mean, in the end, if it runs, is it worth figuring out where that piece went? Or not and better to ignore it?
crimson ferry
I would think authmodes would be more useful for Access Point
jaunty juniper
you can remove authmodes or add something like:
print(" - Auth: {}".format(authmodes[authmode]))
I tried deleting both authmodes and authmode but it fails on line 14 with ValueError: too many values to unpack (expeccted 5)
Hmm ok...
jaunty juniper
no you can't change the tuple in the for
crimson ferry
yeah, or that too
jaunty juniper
it's unpacking each result of the scan
idle owl
Ah ok
jaunty juniper
like that here, somebody's printer:
(b'DIRECT-C6-HP OfficeJet Pro 6970', b'\x80\xe8,\x02\x1c\xc7', 6, -81, 3, False)
(and you see authmode is 3, which is 'WPA2-PSK')
is the 4 a typo too ?
idle owl
Someone's random WiFi scanning with the code I posted above. ```* SunGuest
- Channel: 5
- RSSI: -88
- BSSID: 92:....
- Auth: WPA/WPA2-PSK```
idle owl
Thank you so much! I think I have the right amount of info to take to Limor, and working code either way. Much appreciated!
crimson ferry
you may want to use some caution posting full MAC addresses, they can provide location info
crimson ferry
right, either way
idle owl
um... Can I do a string that uses .format(*bssid) but only displays part of it?
crimson ferry
":".join("%02X" % _ for _ in buf[4:7]) + ":RE:DA:CT"
idle owl
Or does it matter for the code, it's don't paste it into Discord?
crimson ferry
just Discod
idle owl
Oh ok ok ok. That makes more sense. Thank you for the heads up.
crimson ferry
oops, I should fix that code line, it's adapted for something specific
first three bytes is manufacturer, last three bytes is the specific device
jaunty juniper
not sure how you get to a location from the MAC of an access point ๐ค
crimson ferry
there are databases
idle owl
I'm happy to be paranoid about that sort of thing.
crimson ferry
same
jaunty juniper
I mean, in google's vaults maybe ?
crimson ferry
jaunty juniper
like, from phones pinging locations of networks they detect ?
idle owl
@crimson ferry Do you have info on what part is ok to display? Or is it the whole thing together that matters, and showing the first two parts is ok or something?
I need to edit the screenshot for this guide to cover them up.
crimson ferry
it's pretty much ok to display the first three bytes, since that's just manufacturer
idle owl
Keen thank you.
jaunty juniper
oh yeah some app that does some snooping
crimson ferry
Mobile app makers (or their 3rd-party ad "partners") LOVE to have access to wifi to see what's around to pinpoint location even if you have location off (can't do that anymore on iOS from what i understand). Also, retail areas love to watch even client devices as they roam around.
(I just try to stay away from posting any unique identifiers)
idle owl
(I just try to stay away from posting any unique identifiers)
Valid approach.
crimson ferry
.eu has better laws, but the U.S. is the wild west for data aggregators
ok, enough #general-chat
idle owl
Heh. Nah, all good. Thank you for the pointers.
Passed Pylint remotely for the PR. And works running it on the board. Seems like a win to me.
idle owl
Blinka-purple border:
Guess that answers that question. Blinka-purple border it is!
misty garnet
does anyone use VSCode + WSL to do their development and building?
tulip sleet
I guess you could fix the missed overflow problem (the hypothetical "if you some...
Do you want to add it back? If someone is timing something that's really long, it could go that long.
onyx hinge
@tulip sleet so the thought would be, it doesn't matter if in the irq the overflow hasn't happened yet, just so that the old-ticks value gets updated so on a real call after overflow the overflow occurs. however, I'm not sure that's bulletproof, because (say) the pre-overflow value could be 0xffff_fffc one time and then 0xffff_fffd the next "overflow", it would not record an overflow.. if we actually care about that case, more thought is required.
tulip sleet
<@!329766224093249548> so the thought would be, it doesn't matter if _in the irq...
do you mean if you got two overflows in a row (as opposed to the next time around)?
onyx hinge
no, I meant 3 days later
tulip sleet
oh, i see if it was not called at all the next cycle, I see
onyx hinge
the userspace code has to be very careful never to call a time function over 3+ days, or it would work just fine. this would be an "every 6 days, a slim chance" problem.
userspace isn't quite the right term
tulip sleet
I think the current PR is better than the existing code, in any case, because waiting a complete cycle before calling time is much rarer than calling it a lot
so let's merge and you can write a "Very Long Term" issue. Does that make sense?
onyx hinge
I also think it's an improvement, though it still isn't tested on samd21 by me
either just in the normal case or in the overflowing case
tulip sleet
can you easily test it with your setting of the clock reg? I think you are almost all set to do that. That would be great.
I would include this in beta.0 or whatever
onyx hinge
write on the PR for me to do it and I'll try to get to it on monday ๐
