the brew page said something about being able to alias it as make but I didn't read that part for comprehension
#circuitpython-dev
onyx hinge
idle owl
Heh yeah
It's a thing
That I don't know how to do without explicit instruction.
Do I need new make? Uff, probably.
I have an existing PATH somewhere, and it's not in my .bashrc which is where the instructions tell you to add PATH="$(brew --prefix)/opt/make/libexec/gnubin:$PATH" to alias it. I am confuse.
onyx hinge
that's good, because you know you're likely to end up working on the guide text
onyx hinge
the part I didn't say is that you know you're well-suited to writing stuff like that up
onyx hinge
I made an extremely minimal library for CP1203, which has a qwiic board from adafruit that I've owned for many years and not been able to use: https://gist.github.com/jepler/4e1840fe9a37a500d13287870c1f03d7
I didn't get the multitouch stuff to work but I didn't try too hard.
proven garnet
What's more beginner friendly - requiring an input be str or calling str() on inputs to turn them into strings?
random junco
the latter
proven garnet
Context is they provide something to be pasted to a site like PasteBin
crimson ferry
hmm...```py
os.statvfs("/")
(512, 512, 1881, 1665, 1665, 0, 0, 0, 0, 255)
storage.erase_filesystem(extended=True)
[20:23:34.895] Disconnected
[20:23:44.929] Connected
os.statvfs("/")
(1024, 1024, 2348, 2344, 2344, 0, 0, 0, 0, 255)copy half a meg of files to CIRCUITPY
os.statvfs("/")
(1024, 1024, 2348, 2344, 2344, 0, 0, 0, 0, 255)
ah, after a[nother] reset, it updates```py
os.statvfs("/")
(1024, 1024, 2348, 1681, 1681, 0, 0, 0, 0, 255)
Nice feature, <1MB flash is tight
graceful rain
Hi all, apologies for cross-posting a Git problem, but Paul directed me to post here, so here goes:
I'm in Ubuntu 22.04.1 LTS and made some CP changes and trying to commit to my branch, and getting an error I've never seen before:
[INFO] Installing environment for https://github.com/pre-commit/pre-commit-hooks.
[INFO] Once installed this environment will be reused.
[INFO] This may take a few minutes...
An unexpected error has occurred: AssertionError: BUG: expected environment for python to be healthy immediately after install, please open an issue describing your environment
more info:
virtualenv python version did not match created version:
- actual version: <<error retrieving version from /home/gpeyton/.cache/pre-commit/repodk7pays9/py_env-python3/bin/python>>
- expected version: 3.10.6.final.0
Check the log at /home/gpeyton/.cache/pre-commit/pre-commit.log
Any tips? I'm on an almost brand-new install of Ubuntu and clean branch with only a ports folder added. I'm working with ESP32, so same error before and after I export.sh to setup the environment, so it looks unrelated to that.
proven garnet
the latter
Thanks!
onyx hinge
@graceful rain yes this is familiar, but let me see if I can find the chat about it
In ubuntu 22.04 or the analogous Debian release, you may see the error "expected environment for python to be healthy immediately after install" when trying to use pre-commit. To fix this, add this line to your .bashrc or .bash_aliases file, or other shell startup file. Restart your shell as necessary to pick up this setting.
export SETUPTOOLS_USE_DISTUTILS=stdlibThis export must be present before pre-commit sets up its virtualenv environment, which happens the first time you do pre-commit run or you try to push a commit. If the virtualenv is already set up, do pre-commit clean, which removes the existing virtualenv.
See the instructions from the pre-commit project for installation for alternative ways of installing pre-commit.
I think this may still be the best workaround for the problem, you can read more about it near this in old messages to find links to more info about it
graceful rain
wow, it's a long thread. Ok, let me take a look at that article first. Thx!
onyx hinge
yeah it puzzled some of us for days. even when we started finding all the related issues on github it was still unclear. and most of us had installs old enough to not be affected no matter what we tried.
graceful rain
Thanks I think that did it!
onyx hinge
yay! goodnight!
spiral elk
Is the update to esp-idf 5.0 intended to be a pre or post-8.0 thing?
analog bridge
Is the update to esp-idf 5.0 intended to be a pre or post-8.0 thing?
its likely to be post-8.0
proven garnet
Does the QT Py S2 (and possibly other related boards) have an issue connecting to wifi when connected to serial as well? I noticed during my CircuitPythonukiah project I can connect to networks when hooked into the board as a USB, but it works fine if I power the board via the pins.
Also seeing it an another one of my QT Py S2 boards
jaunty juniper
can't connect ?
midnight ember
Are you running 8.0 beta 2 or above on it? Wifi issues should have been fixed with the beta 2 release. I didn't have any issues on the QT Py S3 running an open weathermaps script even during the beta 1 hard reset bug issue. It was the only board I had that was unaffected. It's not an S2 though.
proven garnet
can't connect ?
Yeah, to the network
proven garnet
Are you running 8.0 beta 2 or above on it? Wifi issues should have been fixed wi...
beta.3 actually
But I also saw it on 2 I think
midnight ember
@proven garnet apparently I do also have a qt py s2. Do you have a test script I can run to help confirm?
proven garnet
Not on hand, but see if you can connect a few times to your network using wifi.radio.connect() with something like tio or Mu active
To be fair I also had issues even if the serial connection wasn't working
midnight ember
Oh I have plenty of those snippets on hand. Ok will do in a little while.
proven garnet
I also recently changed routers and quadruple checked my credentials, for what it's worth.
midnight ember
I have a qt py s3, can test on both.
My Wi-Fi’s been pretty good with the feather s2 lately but not perfect either.
Do you have a bug report open for it yet where I can read up on expected behavior?
Or is this still pre-bug report?
proven garnet
Yeah I wanted to check before submitting mostly since I figured it might be a duplicate if I'm hitting it - I don't usually do a lot of wacky core testing so I'm usually last haha
I can submit one today
midnight ember
Adafruit CircuitPython 8.0.0-beta.3 on 2022-10-20; Adafruit QT Py ESP32S2 with ESP32S2
unknown failure 15
Auto-reload is on. Simply save files over USB to run them or enter REPL to disable.
code.py output:
]0;🐍Wi-Fi: off | code.py | 8.0.0-beta.3\]0;Wi-Fi: No IP | code.py | 8.0.0-beta.3\
===============================
Connecting to WiFi...
Connection Error: Unknown failure 15
Retrying in 10 seconds
Connection Error: Unknown failure 15
Retrying in 10 seconds
Connection Error: Unknown failure 15
Retrying in 10 seconds
Connection Error: Unknown failure 15
Retrying in 10 seconds
using my basic get time from online here: https://github.com/DJDevon3/My_Circuit_Python_Projects/tree/main/Circuit Python Tips %26 Tricks
same after hard reset, no joy. qt py s2 is borked in beta 3.
actually here's the small test script that i came up with. works on qt py s3 on beta 1, does not work on qt py s2 on beta 3.
# SPDX-FileCopyrightText: 2022 DJDevon3 for Adafruit Industries
# SPDX-License-Identifier: MIT
# UM ESP32 FeatherS3
import wifi
import socketpool
# Initialize WiFi Pool
pool = socketpool.SocketPool(wifi.radio)
try:
from secrets import secrets
timezone = secrets['timezone']
tz_offset_seconds = int(secrets['timezone_offset'])
except ImportError:
print("Secrets File Import Error")
raise
# Connect to Wi-Fi
print("\n===============================")
print("Connecting to WiFi...")
while not wifi.radio.ipv4_address:
try:
wifi.radio.connect(secrets['ssid'], secrets['password'])
except ConnectionError as e:
print("Connection Error:", e)
print("Retrying in 10 seconds")
# With the 205 Error it never reaches below this point
print("Connected!\n")
that's about as simple as it gets, still throws unknown failure 15
bug confirmed
onyx hinge
@midnight ember do you have any other espressif board to test on? in any case, please file an issue on github.
with the simple example 🙂
midnight ember
im confirming for anecdata
can load beta 3 on the qt py s3 as well?
i haven't installed beta 3 on any of my boards prior to this, most are still on beta 2, been too busy with projects
midnight ember
i've been trying for the past 10 minutes to get the qt py s3 into bootloader to flash to beta 3. it will not go into bootloader mode.
web flasher doesn't work with it either. looks like the COM port is getting locked to the device, very similar if not the same behavior i encountered with the esp32-V2 when i first got that one.
gonna have to use python esptools to flash it to beta 3. anyone running beta 1 on the qt py s3 is gonna be screwed if they don't know how to use esptools.
proven garnet
Thanks @midnight ember for helping to confirm it's not just me!
midnight ember
np, that's a big bug
i should file a bug report for the qt py s3 beta 1 bootloader issue just in case anyone runs into it. it exists so should be documented if it hasn't been already?
analog bridge
@midnight ember do you have CP active on qtpy s3?
midnight ember
yes
analog bridge
are you trying get into bootloader mode pragmatically?
midnight ember
Adafruit CircuitPython 8.0.0-alpha.1 on 2022-06-09; Adafruit QT Py ESP32-S3 no psram with ESP32S3
tried every reset key combination, doubled checked the qt py s3 guide just to be sure it's not me...
oh it's alpha 1
welp that one serves me right lol
analog bridge
import microcontroller
microcontroller.on_next_reset(microcontroller.RunMode.BOOTLOADER)
microcontroller.reset()
does this work
midnight ember
nope, refuses to show featherboot
had to get into safemode to get out of it, whew
analog bridge
featherboot? do you mean to go into UF2 bootloader? For that try, microcontroller.RunMode.UF2
midnight ember
yes thought they were the same mode?
onyx hinge
esp32s2 and s3 have two different bootloaders -- the ROM bootloader which works with esptool, and the optional uf2 bootloader that works with, well, uf2 files.
analog bridge
BOOTLOADER is the default rom dfu bootloader, the one used for serial flashing and UF2, as the name suggests is the uf2 bootloader which makes the drive appear
onyx hinge
additionally i've found that loading anything with esptool usually means the uf2 bootloader won't work anymore, until I re-load it with esptool
midnight ember
yeah that didn't work either, gonna have to nuke it. will do that later i'm about at the end of my day. been up all night pumpkin carving.
this one is on me because it's running an alpha build, will fix that in my own time. doubt it needs a bug report now.
will take another look at it later tonight and get beta 3 on the qt py s3 and go from there.
crimson ferry
@proven garnet @midnight ember this works fine for me on Adafruit CircuitPython 8.0.0-beta.3 on 2022-10-20; Adafruit QT Py ESP32S2 with ESP32S2:```py
import time, wifi, socketpool
from secrets import secrets
pool = socketpool.SocketPool(wifi.radio)
def connect():
try:
wifi.radio.connect(secrets['ssid'], secrets['password'])
except ConnectionError as e:
print("Connection Error:", e)
while True:
print(f"Connecting... ", end="")
wifi.radio.enabled = True
connect()
print(f"Disconnecting... ")
wifi.radio.enabled = False
time.sleep(1)
Cool. Can you get it to the uf2 bootloader? 😆
Oh wait that’s for the s2…
Are you using it on usb?
Not web workflow?
crimson ferry
This puts it in UF2 bootloader: import microcontroller ; microcontroller.on_next_reset(microcontroller.RunMode.UF2) ; microcontroller.reset(). It's on USB. Web Workflow is not enabled.
midnight ember
Weird. How are tekktric and I getting the same bug but you’re not? 2/3.
Different bootloader versions? Pretty sure I’m still on the one it shipped with.
crimson ferry
I don't have a QT Py S3, but I have. Feather S3. An S3 is an S3 imo. Same code works on the S3.
midnight ember
The bug tekktric had was on the qt py s2.
crimson ferry
right, that's the first thing I posted. I suspect the issue isn't inherently with beta.3
midnight ember
Oh you’re saying I don’t have to bother testing the qt py s3.
crimson ferry
well, you can, but without anything using pins, there shouldn't be a difference?
(there are flash and ram differences, but this simple code shouldn't touch that)
midnight ember
Thought it was a power delivery difference using pins vs usb..
Not a connected sensor or something.
crimson ferry
could be a power thing, or the flash needs erasing, or there's a network issue, or...
I have 99 problems, and the network is 98 of them
(I kid, the network is rarely my issue)
midnight ember
My feathers work fine but I’m still on beta 2. Don’t have boards I can free up for a thorough testing of beta 3 yet.
Seems like beta 3 definitely needs to be put through paces.
crimson ferry
I've switched over a bunch to beta.3, more solid than previous
midnight ember
Developers, I love you, but this cycle of a new release per week I can’t keep up with lol. Going too fast. 😂 good problem to have really.
crimson ferry
there's always 7.3.3 😉
I still have some long-running boards on 6.x o_O
a couple in hard-to-reach places... if it ain't broke...
midnight ember
Well a couple of my projects were broken so I had to. The social media tracker box for example. That’s stable now so I lost that board for beta testing as an in production project.
Same with another s2 & s3 I’m going to use for Halloween projects. Right now, this week in particular is a bad week to need beta testing as my boards are being used in Halloween projects.
I imagine at least a few other people are probably in the same boat.
crimson ferry
(I did get 3 Connection Error: Unknown failure among 500 connections: (1) 2; (1) 4; (1) 200)
manic glacierBOT
CircuitPython version
Current
Code/REPL
not relevant
Behavior
Description
On the README page https://github.com/adafruit/circuitpython, README.rst, under Documentation, I clicked on the link Read The Docs (http://circuitpython.readthedocs.io/en...
lone axle
is there some "trick" to download the builds from artifacts in github? Beyond clicking the link of the artifact name on pages like this https://github.com/adafruit/circuitpython/actions/runs/3322093441
onyx hinge
I think you have to be logged into github
lone axle
clicking the artifact name makes a blue progress bar at the top of the page which goes very slowly and eventually gets all the way to the right but nothing else happens.
onyx hinge
weird
jaunty juniper
I got a pink unicorn clicking on your link
lone axle
I am logged in. Perhaps their system is having trouble today. I just downloaded one last week and I don't think I did anything differently. I don't do it often enough to be confident I have the process right though 😅
jaunty juniper
there might be github issues right now ?
crimson ferry
github issues in general
lone axle
thank you
crimson ferry
not showing in https://www.githubstatus.com though
jaunty juniper
same issue with the blue bar but I did get the zip by middle clicking the zip link/copy and pasting it
onyx hinge
is anyone else getting a 403 forbidden at https://docs.circuitpython.org/en/latest/shared-bindings/_bleio/index.html#bleio.RoleError ? I'm not but a user reports it above.
crimson ferry
Should watchdog be triggering on a long file read?
jaunty juniper
Should watchdog be triggering on a long file read?
like a crash/reset ? or the watchdog module watchdog setup and triggering because the read is too long ?
by the way I was thinking of your issue with loading a large dict, do you have the option to output it as a python file ? (and then mpy-cross on it)
or is it loaded through wifi from an external source you don't control
crimson ferry
I'm pretty sure the watchdog module is triggering because the file read is taking longer than the watchdog threshold
I have complete control over the large data set... I've tried large CSV, large JSON, breaking it up into smaller files, etc. The issue is trying to get all the data into memory to avoid disk accesses.
jaunty juniper
and it's a dictionary structure, so loading it into a bytearray is impractical
crimson ferry
large JSON on a Feather S3 TFT: MemoryError: memory allocation failed, allocating 292856 bytes Seems to work better when I break it up into smaller files, but then I get a lot of it, and a smaller memory allocation error, even when I load the smaller files largest first.
I was thinking of trying two separate lists, since the keys are already sorted
but I'm really not sure if that would buy anything
I also might try integer keys. Will also see how bad caching from some efficient file search is... pain to start up, but then should get better as more is in cache.
jaunty juniper
I'm curious how a python dict in an mpy file would perform here
crimson ferry
I could try that
crimson ferry
I get a memory allocation error on my Mac using mpy-cross o_O
crimson ferry
just that file (macOS 12.6) latest mpy-cross, just downloaded
yeah, that shocked me
jaunty juniper
I'm generating a python file with something like that and getting a memory error with mpy-crossed when it gets around 2600
print("data = {")
for x in range(2500):
print(f""" "key-{x}": "Some value that is not {x}",""")
print("}")
so, that doesn't help
crimson ferry
that's a new one
jaunty juniper
I think it's file size
crimson ferry
memory error trying to import the .py dict, as expected
I could go to an 8MB RAM board, but I really want either the TFT for compact display, or QT Py for the antenna choices. Also, this is a fun challenge 🙂
waits patiently for N8R8 QT Py S3 with u.FL ;-)
onyx hinge
I love it when an unrelated realization unblocks something. It is sloowwwwwwwwwwwwwww but I can serve https from my picow now.
the actual problem was that blocking/non-blocking bug in the httpserver module! the rest worked
thorny dove
@onyx hinge & @jaunty juniper as the BOT already signaled, I managed (also) to create a build with DEBUG=1. I think the serial output results are interesting, beside the fact - that also @jaunty juniper has seen in his build with DEBUG=1, that the WDT had reset the processor, this time the serial output gives more information, various Errors in red color like (3115) "HW TIMER NEVER INIT ERROR" (4x) and (5035) "phy_init: esp_phy_load_cal_data_from_nvs: NVS has not been initialized. Call nvs_flash_init before starting WiFi/BT". Also a warning: (5045) "phy_init: failed to load RF calibration data (0x1101), falling back to full calibration"
jaunty juniper
I get a freeze when I try to connect with the FeatherS3 right now >_>
D (226160) wifi:new_bss=0x3fca9c70, cur_bss=0, new_chan=<8,0>, cur_chan=1
D (226170) wifi:filter: set rx policy=5
I (226170) wifi:new:<8,0>, old:<1,0>, ap:<255,255>, sta:<8,0>, prof:1
D (226180) wifi:connect_op: status=0, auth=5, cipher=3
D (226180) gdma: new group (0) at 0x3fcb8240
D (226180) gdma: new pair (0,0) at 0x3fcb827c
D (226190) gdma: new tx channel (0,0) at 0x3fcb80f0
D (226190) gdma: new rx channel (0,0) at 0x3fcb829c
Guru Meditation Error: Core 0 panic'ed (LoadProhibited). Exception was unhandled.
Core 0 register dump:
PC : 0x40056ea0 PS : 0x00060130 A0 : 0x8002f1fd A1 : 0x3fcee000
A2 : 0x00000001 A3 : 0x00000000 A4 : 0x00000000 A5 : 0x3fcee0a3
A6 : 0x4002eb04 A7 : 0x00000000 A8 : 0x3fcf0ec7 A9 : 0x3fcf0ec8
A10 : 0x3fceea80 A11 : 0x00000020 A12 : 0x3fcf0c44 A13 : 0x3fcee0a3
A14 : 0x3fcee079 A15 : 0x00ffffff SAR : 0x0000000f EXCCAUSE: 0x0000001c
EXCVADDR: 0x00000001 LBEG : 0x400570e8 LEND : 0x400570f3 LCOUNT : 0x00000000
Backtrace: 0x40056e9d:0x3fcee000 |<-CORRUPTED
ELF file SHA256: 77c3d51de8b98c96
CPU halted.
but that might be only in the debug build
lavish saffron
Hi, when adding typing to a library and the doc-string has :param int timeout: ... but all the math is with time.monotonic which is float math, should I respect the doc-string because Python will cast the int as a float, or should I change the doc-string to :param float timeout ? Thx
proven garnet
Hi, when adding typing to a library and the doc-string has `:param int timeout: ...
There can be a mismatch. Both should follow the actual use case. In the case of timeouts using monotonic, you can use float, because all int are float in the numeric tower.
The type annotation is also about the promise to the user for what can be used, so in this case I would change the docstring to float because it's valid to use it in this case.
manic glacierBOT
lone sandal
midnight ember
Got the QT Py S3 flashed with beta 3 this morning. Connects to wifi np.
even after soft reboots so no hard-reset bug either.
immediately swapped in the QT Py S2 with the same code ```py
Connecting to WiFi...
Connection Error: Unknown failure 15
Retrying in 10 seconds
Connection Error: Unknown failure 15
Retrying in 10 seconds
Hi, how should I type these? Thx
`
def enter(self):
return self
def exit(self, exc_type, exc_val, exc_tb) -> None:
...
`
stuck elbow
why would you? it's not like you are ever going to call them directly
tulip sleet
Hi, how should I type these? Thx
`
def __enter__(self):
return self
def __...
is this in libraries or shared-bindings? You can look through existing libraries to find similar examples. The easiest way to do that is to clone the Bundle repo and get all the submodules, which are all the libraries in the bundle. Then you can search through all the libraries at once.
lavish saffron
is this in libraries or shared-bindings? You can look through existing libraries...
It's in the socket class of the Adafruit_CircuitPython_Wiznet5k library. I'll clone the Bundle, an excellent suggestion, thanks.
jaunty juniper
yeah it's in many other libraries
proven garnet
immediately swapped in the QT Py S2 with the same code ```py
Connecting to WiFi....
Wacky, do you want to submit the issue since you debugged it more?
proven garnet
Hi, how should I type these? Thx
`
def __enter__(self):
return self
def __...
You can type them if you want, but I wouldn't worry about it like @stuck elbow said if you don't want to since they aren't accessed directly. If you want though, the motor library has __enter__() and __exit__() typed:
Most of the "dunder" (__abc__()) methods have already been typed in at least one other library I think, as has been mentioned.
lavish saffron
My preference is to leave them untyped
stuck elbow
I wonder why mypy doesn't type them implicitly, after all it should know what their types are
proven garnet
I think IntelliSense just gives you the expected outputs, so I'm also surprised mypy doesn't do that.
I can't think of a dunder method that has a "variable" return
Though jepler did mention __init__() could return something other than None that could be accessed... for reasons?
Also, looking for general feedback on this new pastebin library. I tried to keep it extremely simple, just init and dump text, but didn't know if it was beginner-friendly or missing anything: https://github.com/tekktrik/Adafruit_CircuitPython_Pastebin
Wanted something like a uniform API when the services vary in what knobs can be turned when uploading content.
midnight ember
@proven garnet sure can do
analog bridge
@onyx hinge You have marked the coproc PR as draft, are more changes in the pipeline?
onyx hinge
<@426200547024961539> You have marked the `coproc` PR as draft, are more changes...
I had intended to test it on an esp32s2 but didn't find the time yet. I marked it as ready for review just now
proven garnet
@tulip sleet I have an old PR that you think you had said you were going to take a look at later, just wondering if I should close it since its old or if things are still in flux. Or I can retag librarians if you're just swamped: https://github.com/adafruit/Adafruit_CircuitPython_ESP32SPI/pull/160
Remembered it because you recently made an issue in HTTPServer about WSGI stuff and didn't know if that also had an effect.
tulip sleet
<@329766224093249548> I have an old PR that you think you had said you were goin...
I'm sorry I haven't acted on this. I have an incompatible potential PR for adafruit_requests in this branch: https://github.com/dhalbert/Adafruit_CircuitPython_Requests/tree/change-exceptions which I haven't finished. I didn't know whether it would intersect with the PR above. I looked briefly at your PR and it could probably stand on its own. I would really like to regularize the ESP32SPI and adafruit_requests API's, remove the all the "legacy" stuff that is a legacy of old API's in ESP32SPI (and has leaked into various other libraries) but I haven't had time to do that. The exeptions should be a lot more like the CPython ones, though there are not always obvious mappings because what's happening at the low level is different.
I was just saying in an internal staff meeting a couple of days ago that it would be great if someone with network experience could take this rather "strategic" task on. I don't know if you feel you would like to or not.
tulip sleet
Some of the exception cleanup may have been done already in some other adafruit_requests PR's (or just one) this year -- I have to check
proven garnet
I'm a big fan of the WSGI and network PRs, though not sure if the "network experienced" part applies haha. Definitely would be open to it though.
Definitely good at hitting exceptions though
I can at least start looking into it
And no worries, thanks for the context!
Are you talking about the old legacy requests stuff like set_socket() as well?
idle owl
@tulip sleet Do you still want me to test this? I see Jeff approved it.
tulip sleet
Here is the notes document for next Monday’s CircuitPython Weekly meeting. It is at the normal time of 11am Pacific / 2pm Eastern here on Discord. Everyone is encouraged to attend! Please add your hug reports and status updates even if you’ll be attending the meeting - it’s super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and we’ll read them off during the meeting. Hope to see you there! <@&356864093652516868>
https://docs.google.com/document/d/1a4Io4HgOm4ZJtsJRULOcBKeIPnKR7w8GChuI3TzJ0ho/edit?usp=sharing
tulip sleet
<@329766224093249548> Do you still want me to test this? I see Jeff approved it.
If you have a setup to test it yes, that would be great. Jeff checked over the code logic but did not test it. I tried your example with value=True and value=False, and it seems to do what I would expect now, including printing messages at the right time, but good for you to check too.
idle owl
If you have a setup to test it yes, that would be great. Jeff checked over the c...
I think I'm still set up somewhere. My desk is such a mess, I'm fairly certain the board is still in here somewhere. I'll take a look today or tomorrow.
tulip sleet
I think I'm still set up somewhere. My desk is such a mess, I'm fairly certain t...
I might merge it sooner if I want beta.4 more quickly, but a test would still be good. It's better, I just want to make sure it's right now.
idle owl
I might merge it sooner if I want beta.4 more quickly, but a test would still be...
Sounds good. I have no objections to testing it in beta.4 as well. Merge whenever works for you and I'll test it accordingly.
idle owl
@proven garnet I accepted that repo transfer. But the repo isn't showing up in my list. Can you check if it's still yours? And if so, send me a new link? The existing one "expired" because I already clicked it, I assume, since it hasn't been long enough for it to timeout expire.
It always says "This may take a few minutes" but it never does. Also it's been a few minutes at this point.
proven garnet
<@135415127879974912> I accepted that repo transfer. But the repo isn't showing ...
Huh, weird, I do still have it
idle owl
Ok, please try again to transfer it
idle owl
Clicked on my end. Leaving the page alone for a few minutes this time in case that's the problem.
crimson ferry
thanks Dan, tags and titles are really useful for future code archaeologists 😉
jaunty juniper
@midnight ember the revision is the letter in a circle
midnight ember
ahh yes it's a C
tulip sleet
@analog bridge so my flash-size-mode-freq PR may be all wrong. I thought the builds picked up that info from sdkconfig when they built the .bin's, but you've shown otherwise. If not, then we really need to audit all the settings in the mpconfigboard.mk files, as people frequently copy these blindly.
idle owl
@proven garnet Got it this time!
proven garnet
Huge success!
idle owl
And transferred to Adafruit.
proven garnet
Thanks @idle owl!
ornate breach
@idle owl I love the scorpion, did you have any hand in its creation? 🙂
idle owl
Alas, I cannot claim that.
It was probably Phil B (PaintYourDragon), as he does most of our silk stuff, but it could have also been our graphics person involved as well.
ornate breach
Ah well, it looks cool. Reminds me what I feel like a scorpion should look like
In the words of that obscure SpongeBob character: “BIG. MEATY. CLAWS”
ornate breach
This being compared to the claws many scorpions have which tend to be kind of small
idle owl
I've only ever seen one personally in Italy. It was in the bathroom of our apartment and scared the absolute ... out of my mum. I had to go deal with it. Then the next year around xmas, she found one that looked a bit like it in resin, and bought it for me. It's lived in my bathroom on the counter since.
ornate breach
I've only ever seen one personally in Italy. It was in the bathroom of our apart...
What a momento 😳
idle owl
Hehehe.
stuck elbow
<@330227457296957440> I love the scorpion, did you have any hand in its creation...
what scorpion?
ornate breach
what scorpion?
stuck elbow
it lost two legs, poor thing
idle owl
It was our graphics person Bruce! I asked around 🙂
idle owl
it lost two legs, poor thing
Um... no it didn't?
I had to google, but the grabbyhands are their 7th and 8th legs, at least at first glance.
ornate breach
Wait, don’t they have 10 appendages?
stuck elbow
scorpions have 8 walking legs, just like their cousins spiders
idle owl
ornate breach
Google says 6-12
stuck elbow
the claws are the equivalent of pedipalps in spiders
ornate breach
idle owl
The thing I know about them is that the smaller and clearer they are, the more poisonous they are. The large black ones are basically the least poisonous.
ornate breach
The ones you don’t see and live by are the safest
idle owl
the claws are the equivalent of pedipalps in spiders
Interesting.
stuck elbow
scorpions are basically spiders with opposable thumbs
idle owl
That description is basically terrifying.
ornate breach
This will be the image you see for the rest of your life
Anyway, scorpions 6-12 legs and two thumbs way up
Depending on the specific type of scorpion
stuck elbow
Anyway, scorpions 6-12 legs and two thumbs way up
appendages, depending on how you count
ember iris
Arms are legs if you walk on them
So that's how you get a leg up in life
stuck elbow
I think I still trust Wikipedia more than Wildlife Journal Junior
stuck elbow
It’s PBS, government funded
if we can't trust the governments of the world, whom can we?
ornate breach
It was our graphics person Bruce! I asked around 🙂
Do pass along that it looks great
idle owl
Do pass along that it looks great
PT said "If someone hates it or it's offensive, then I did it. I'll apologise and make sure it gets updated." I piped up that (you) loved it. He said, "Oh, ok good." 😄
ornate breach
I’m glad I’m the opinion of reason there hahaha
stuck elbow
it's cute
ornate breach
PT said "If someone hates it or it's offensive, then I did it. I'll apologise an...
Did the “Blinka squashing bugs” graphic ever make the chopping block?
Not sure I’m using the right analogy
idle owl
Did the “Blinka squashing bugs” graphic ever make the chopping block?
No! But to be honest, it's still on my list. So it hasn't been approved or denied yet. I am still working on what it might look like so I could provide Bruce with a sketch (which he would end up doing much better anyway, but still).
The way you worded it, it sounds like you're asking if it was denied.
ornate breach
Opposite, approved lol
idle owl
Hah
ornate breach
Anyway, I was thinking of Blinka racing after little green circuit bugs or something like that
Or maybe holding a fly swatter
ornate breach
My other thought was blinka controlling a circuitpython powered squashing robot lol
idle owl
Hah! That one would be beyond my sketching skills. But I bet Bruce would still produce excellence.
idle owl
That was all him. I told him what you said, and that's what he came back with.
ornate breach
Yeah, it was great
lavish saffron
Hi, a typing query. The function below returns the buffer when it is called with a buffer and returns a bytearray when called with None. Should I install typing-extensions and use the @overload decorator or type the return as Union[WriteableBuffer, bytearray] or something else?
def read(buffer: Optional[WriteableBuffer] = None):
...
if buffer is None:
self._rxbuf = bytearray(length)
bus_device.readinto(self._rxbuf) # pylint: disable=no-member
return self._rxbuf
bus_device.readinto(buffer, end=length) # pylint: disable=no-member
return buffer
@analog bridge HI, I see you did an excellent job with solving issue #7060. I am stuck with a boot-loop on an Adafruit ESP32-S3-Box. Danh, Jepler and Neradoc have paid some attention to it. Danh is very busy with other bugs. So, in this moment there is no activity. No advance into solving the bug. Could you please have a look into issue # 7093? Title: 'esp32-s3-box with cpy 8.0.0-b.2 loops in reboot as soon as .env file is uploaded.' I am now testing with the latest: cpy 8.0.0-b.3-12 however the problem continues. It also happens when the .env is not on the drive, when I try to connect a WiFi (extender) AP here in the room, either from a CPY script or from REPL. The WDT resets the PRO CPU and it resets the AP CPU. All other WiFi 'steps' e.g. Networks scan are executed without problem.
analog bridge
<@747673663419187281> HI, I see you did an excellent job with solving issue #706...
Hi! @thorny dove, I don't have a esp32s3 board handy so I won't be able to run any tests.
I did a DEBUG=1 build with current main and have attached the bin and uf2 files for espressif_esp32s3_box.
I might be able to provide some insights if you can get the debug output.
from what I read in the issue comments, simply trying to connect to WiFi results in a crash? Can you test this with the following:
import wifi
wifi.radio.connect("ssid", "password")
@analog bridge Hi, thank you for your reply. Yes, the command wifi.radio.connect("ssid", "password") does result in a crash. I just finished creating a build, uising esp-idf v 5.0; next flashed the tinyuf2 bootloader; finally copied the firmware.uf2. A simple WiFi test resulted in a 'crash' (reset). See the following extract of the serial output: ```
(105071) wifi:first chan=1
D (105071) wifi:handoff_cb: status=0
D (105071) wifi:ap found, mac=e4:c3:2a:e3:14:31
D (105081) wifi:new_bss=0x3fcaa890, cur_bss=0, new_chan=<9,0>, cur_chan=1
D (105081) wifi:filter: set rx policy=5
I (105091) wifi:new:<9,0>, old:<1,0>, ap:<255,255>, sta:<9,0>, prof:1
D (105091) wifi:connect_op: status=0, auth=5, cipher=3
ESP-ROM:esp32s3-20210327
Build:Mar 27 2021
rst:0x8 (TG1WDT_SYS_RST),boot:0xa (SPI_FAST_FLASH_BOOT)
Saved PC:0x40374200
SPIWP:0xee
mode:DIO, clock div:1
load:0x3fcd0108,len:0x15e8
load:0x403b6000,len:0xbf8
load:0x403ba000,len:0x31c8
entry 0x403b621c
I (29) boot: ESP-IDF v4.4.2-378-g9269a536ac 2nd stage bootloader
I (29) boot: compile time 11:36:39
I (29) boot: chip revision: 0
I (32) qio_mode: Enabling default flash chip QIO
I (37) boot.esp32s3: Boot SPI Speed : 80MHz
I (42) boot.esp32s3: SPI Mode : QIO
I (47) boot.esp32s3: SPI Flash Size : 16MB
Is this the output from the build files I provided?
thorny dove
Yes. I go now to upload the complete serial output to Github > Adafruit > Circuitpython > Issues #7093
analog bridge
Yes. I go now to upload the complete serial output to Github > Adafruit > Circui...
I was hoping for a backtrace in the debug output but it seems like the board resets without printing it.
This is similar to what Neradoc noticed in https://github.com/adafruit/circuitpython/issues/7093#issuecomment-1287941944
Sorry, I am afraid I won't be able to provide much insight without testing on a board myself and no backtrace.
The issue has 8.0.0 milestone so danh will likely get back to it soon, he did mention about updating the idf to a newer commit in release/v4.4 branch.
thorny dove
OK, thank you anyway for trying to help. I much appreciated that! In another test yesterday the serial output clearly stated that it was the internal Watchdog timer that did reset the CPU's.
Here is that part: I (7905) wifi:new:<9,0>, old:<1,0>, ap:<255,255>, sta:<9,0>, prof:1 D (7905) wifi:connect_op: status=0, auth=5, cipher=3 # | <<<===== WiFi Connect command ESP-ROM:esp32s3-20210327 Build:Mar 27 2021 rst:0x8 (TG1WDT_SYS_RST),boot:0xa (SPI_FAST_FLASH_BOOT) # | <<<===== MOMENT OF RESET BY WATCHDOG TIMER Saved PC:0x40374200 SPIWP:0xee mode:DIO, clock div:1 load:0x3fcd0108,len:0x15e8 load:0x403b6000,len:0xbf8 load:0x403ba000,len:0x31c8 entry 0x403b621c I (29) boot: ESP-IDF v4.4.2-378-g9269a536ac 2nd stage bootloader I (29) boot: compile time 11:36:39 I (29) boot: chip revision: 0 I (32) qio_mode: Enabling default flash chip QIO I (37) boot.esp32s3: Boot SPI Speed : 80MHz I (42) boot.esp32s3: SPI Mode : QIO I (47) boot.esp32s3: SPI Flash Size : 16MB W (52) boot.esp32s3: PRO CPU has been reset by WDT. # | <<<======== WATCHDOG TIMER RESETS BOARD W (57) boot.esp32s3: APP CPU has been reset by WDT. # | <<<======== IDEM I (63) boot: Enabling RNG early entropy source... I (68) boot: Partition Table: I (72) boot: ## Label Usage Type ST Offset Length I (79) boot: 0 nvs WiFi data 01 02 00009000 00005000 I (87) boot: 1 otadata OTA data 01 00 0000e000 00002000 I (94) boot: 2 ota_0 OTA app 00 10 00010000 00200000 I (102) boot: 3 ota_1 OTA app 00 11 00210000 00200000 I (109) boot: 4 uf2 factory app 00 00 00410000 00040000 I (117) boot: 5 ffat Unknown data 01 81 00450000 00bb0000 I (124) boot: End of partition table
@analog bridge What if I would build a version of tinyuf2 with ESP-IDF v5.0 ?
analog bridge
<@747673663419187281> What if I would build a version of tinyuf2 with ESP-IDF v5...
tinyuf2 or circuitpython? I have a PR which updates CP to idf-v5.0 https://github.com/adafruit/circuitpython/pull/6913
thorny dove
I was thinking to make a build of tinyuf2 with ESP-IDF V5.0
analog bridge
I created a build with of CP with idf at the latest commit in release/v4.4 branch plus CP customizations
thorny dove
Ok
analog bridge
Can you test this:
thorny dove
Thank you
analog bridge
I was thinking to make a build of tinyuf2 with ESP-IDF V5.0
tinyUF2 shouldn't affect this issue, unless you have noticed that a bin works fine and uf2 dosen't
thorny dove
OK
I reflashed the bootloader then copied your file to it (also V8.0.0-beta.3-20-...) . Same result: Instantaneous reboot
Here is the serial output from the period around the WDT reset: ```
D (8117) wifi:first chan=1
D (8117) wifi:handoff_cb: status=0
D (8117) wifi:ap found, mac=e4:c3:2a:e3:14:31
D (8117) wifi:new_bss=0x3fcaa8e0, cur_bss=0, new_chan=<9,0>, cur_chan=1
D (8127) wifi:filter: set rx policy=5
I (8127) wifi:new:<9,0>, old:<1,0>, ap:<255,255>, sta:<9,0>, prof:1
D (8137) wifi:connect_op: status=0, auth=5, cipher=3
ESP-ROM:esp32s3-20210327
Build:Mar 27 2021
rst:0x8 (TG1WDT_SYS_RST),boot:0xa (SPI_FAST_FLASH_BOOT)
Saved PC:0x403799cf
SPIWP:0xee
mode:DIO, clock div:1
load:0x3fcd0108,len:0x15e8
load:0x403b6000,len:0xbf8
load:0x403ba000,len:0x31c8
entry 0x403b621c
I (29) boot: ESP-IDF v4.4.2-378-g9269a536ac 2nd stage bootloader
I (29) boot: compile time 11:36:39
I (29) boot: chip revision: 0
I (32) qio_mode: Enabling default flash chip QIO
I (37) boot.esp32s3: Boot SPI Speed : 80MHz
I (42) boot.esp32s3: SPI Mode : QIO
I (47) boot.esp32s3: SPI Flash Size : 16MB
W (52) boot.esp32s3: PRO CPU has been reset by WDT.
W (57) boot.esp32s3: APP CPU has been reset by WDT.
I (63) boot: Enabling RNG early entropy source...
I (68) boot: Partition Table:
I (72) boot: ## Label Usage Type ST Offset Length
I (79) boot: 0 nvs WiFi data 01 02 00009000 00005000
I (87) boot: 1 otadata OTA data 01 00 0000e000 00002000
I (94) boot: 2 ota_0 OTA app 00 10 00010000 00200000
I (102) boot: 3 ota_1 OTA app 00 11 00210000 00200000
I (109) boot: 4 uf2 factory app 00 00 00410000 00040000
I (117) boot: 5 ffat Unknown data 01 81 00450000 00bb0000
I (124) boot: End of partition table
proven garnet
@lavish saffron I wouldn't use the @overload decorator since CircuitPython doesn't support it's use. I would use a Union, and I also think this is a good use case for TypeVar
To specify that the exact type of input is the exact type of output when that happens.
overload decorator would apply, but since neither typing nor typing-extensions exists in CircuitPython, we can't do it and the same trick we're using to type annotate with those modules wouldn't work for a decorator.
lavish saffron
To specify that the exact type of input is the exact type of output when that ha...
Thx. I’ll have a look at TypeVar 
deft jay
@tulip sleet do I mark this closed now that we know it runs on beta 2 or...it still doesn't work on beta 3 though.
tulip sleet
<@329766224093249548> do I mark this closed now that we know it runs on beta 2 o...
No, it means there was a regression from beta.2 to beta.3, so it's a new bug
onyx hinge
@tulip sleet huh now I'm having a "how did it ever work" moment. reverting e590d27bf8990f2163ee50e271bc5ce18ee3d956 fixes it, but the code before that was nonsense. Ticks are intended to be micropython small integer objects but in CP we used regular C integers instead ... look at these two implementations of ticks():STATIC mp_obj_t ticks(void) { return MP_OBJ_NEW_SMALL_INT(mp_hal_ticks_ms() & (MICROPY_PY_UTIME_TICKS_PERIOD - 1)); } and #define ticks() supervisor_ticks_ms()
tulip sleet
so maybe we diverted with micropython about the semantics of ticks()??
is this difference particular to the RP2040 impl?
I have this issue on my mind sometimes: https://github.com/adafruit/circuitpython/issues/6452
onyx hinge
oh, supervisor_ticks_ms IS an object after all
that at least satisfies my intuition that it should have been a compile error 😕
onyx hinge
I'd swear up and down that I did test on hardware, but obviously not. 😦 😭
deft jay
yeah, I was thinking, "why isn't this working" until it started working and then I thought "why is this working????"
@onyx hinge is your comment in the github thread ```pyimport asyncio
async def main():
print("in async func")
print("before asyncio run")
asyncio.run(main())
print("after asyncio run")```
in regards to my shoddy code or in regards to something in the backend?
onyx hinge
<@426200547024961539> is your comment in the github thread ```pyimport asyncio
...
that's a small program that seemed to reproduce the problem you're encountering. Not intended to be any comment on the quality of your code.
tulip sleet
@thorny dove I have a version for you to try. Do you need a .bin or a .uf2? (I could not get the bootloader to come up after the first time I used it, so I have been debugging with .bin files)
thorny dove
Thank you. I use espressif download tool V3.9.3. And I have the tinyuf2 bootloader, also OK so you can send me an .uf2
tulip sleet
this works with the standard Adafruit Internet test program that is in all the guides. It is similar to your test program, but the setup is simpler. Also it doesn't do the __init__() call that was in your program. I that is irrelevant to the problem. The problem was incorrect settings for use of the PSRAM (SPIRAM) for the ESP32-S3 Box.
thorny dove
OK
tulip sleet
it should work with your test program too;
thorny dove
It works! Resolved google address: '142.250.200.110'
Ping google.com [142.250.200.110]: 25 ms
Exiting...
]0;🐍192.168.1.112 | Done | 8.0.0-beta.3-20-g2bba6a68e-dirty\
bootlaoader
It works! Resolved google address: '142.250.200.110'
Ping google.com [142.250.200.110]: 25 ms
Exiting...
]0;🐍192.168.1.112 | Done | 8.0.0-beta.3-20-g2bba6a68e-dirty\
tulip sleet
OK! I will make a PR for this and it should be in beta.4 and later. In the meantime you can use that one.
thanks for testing!
thorny dove
Thank you very much !!!
tulip sleet
sure - it was confusing
thorny dove
Yeah. And I threw in a lot of noise. Sorry.
tulip sleet
@onyx hinge should CIRCUITPY_ESP32_CAMERA be turned off on all S3 boards? It was not turned off on ESP32-S3-BOX
thorny dove
@tulip sleet in a later moment I would like to learn what you've changed to get rid of that bug.
tulip sleet
it will be in the PR. The original SPIRAM settings were copied from elsewhere and I think they were just wrong in various ways. I started from a minimal set and just kept adding a few related ones at a time until it worked.
silver tapir
Heads up. Don't click that Upgrade to MacOS Ventura.
https://twitter.com/aallan/status/1585406275781038085
Twitter
thorny dove
OK
tulip sleet
Heads up. Don't click that Upgrade to MacOS Ventura.
https://twitter.com/aallan/...
if you just use cp it works, but the FInder throws an error
this is not the first time a macOS upgrade has broken something about small drive support
😕
@onyx hinge looks ok for S3 after I looked at the PR OP.
onyx hinge
@tulip sleet there aren't enough pins on the 'box so disabling it would be fine too afaik
But it should work on esp32/s2/s3 equally and the point of the new component was to make it work on esp32s3 generally
deft jay
that's a small program that seemed to reproduce the problem you're encountering....
aah, understood, was not at all offended at any remark about my code, I'm very much so learning about these concepts still and just wanted to convey my level of non-expertise. Was asking if the error was caused by my code or a genuine bug. From skimming this it looks like it was a backend bug if I'm not mistaken. Anyway, thank you for looking into it! Sorry to clutter the -dev channel with this chatter.
onyx hinge
@tulip sleet I noticed that the inner loop of TileGrid's pixel generating function uses division and modulo, which are not a single instruction (let alone a fast one) on Cortex M0s like rp2040. I wonder if there's some way to get rid of the division...```
473 int16_t local_y = input_pixel.y / self->absolute_transform->scale;
0x100856a8 <+444>: ldrb r1, [r5, #6]
0x100856aa <+446>: bl 0x1009e7f0 <divmod_s32s32>
are there fast paths in the division routine for / 1 and divide by powers of 2? There might be.
onyx hinge
oh within divmod_s32s32? I duno
tulip sleet
is that code open source?
i was maybe misremembering about the RP2040's ROM fp routines; is int division in there too?
otherwise it's just in libc, I guess
onyx hinge
0x1009e7f0 <+0>: ldr r2, [pc, #68] ; (0x1009e838 <divmod_s32s32_savestate+24>)
0x1009e7f2 <+2>: ldr r3, [r2, #120] ; 0x78
0x1009e7f4 <+4>: lsrs r3, r3, #2
0x1009e7f6 <+6>: bcs.n 0x1009e820 <divmod_s32s32_savestate>
End of assembler dump.
(gdb) disas divmod_s32s32_savestate
Dump of assembler code for function divmod_s32s32_savestate:
0x1009e820 <+0>: push {r4, r5, r6, r7, lr}
0x1009e822 <+2>: ldr r4, [r2, #96] ; 0x60
0x1009e824 <+4>: ldr r5, [r2, #100] ; 0x64
0x1009e826 <+6>: ldr r7, [r2, #116] ; 0x74
0x1009e828 <+8>: ldr r6, [r2, #112] ; 0x70
0x1009e82a <+10>: bl 0x1009e7f8 <divmod_s32s32_unsafe>
0x1009e82e <+14>: str r4, [r2, #96] ; 0x60
0x1009e830 <+16>: str r5, [r2, #100] ; 0x64
0x1009e832 <+18>: str r7, [r2, #116] ; 0x74
0x1009e834 <+20>: str r6, [r2, #112] ; 0x70
0x1009e836 <+22>: pop {r4, r5, r6, r7, pc}
0x1009e838 <+24>: movs r0, r0
0x1009e83a <+26>: beq.n 0x1009e83e <divmod_u32u32+2>
End of assembler dump.
(gdb) disas divmod_s32s32_unsafe
Dump of assembler code for function divmod_s32s32_unsafe:
0x1009e7f8 <+0>: str r0, [r2, #104] ; 0x68
0x1009e7fa <+2>: str r1, [r2, #108] ; 0x6c
0x1009e7fc <+4>: cmp r1, #0
0x1009e7fe <+6>: beq.n 0x1009e80c <divmod_s32s32_unsafe+20>
0x1009e800 <+8>: b.n 0x1009e802 <divmod_s32s32_unsafe+10>
0x1009e802 <+10>: b.n 0x1009e804 <divmod_s32s32_unsafe+12>
0x1009e804 <+12>: b.n 0x1009e806 <divmod_s32s32_unsafe+14>
0x1009e806 <+14>: ldr r1, [r2, #116] ; 0x74
0x1009e808 <+16>: ldr r0, [r2, #112] ; 0x70
0x1009e80a <+18>: bx lr
0x1009e80c <+20>: push {r2, lr}
0x1009e80e <+22>: movs r1, #128 ; 0x80
0x1009e810 <+24>: lsls r1, r1, #24
0x1009e812 <+26>: asrs r2, r0, #31
0x1009e814 <+28>: eors r1, r2
0x1009e816 <+30>: cmp r0, #0
0x1009e818 <+32>: beq.n 0x1009e81c <divmod_s32s32_unsafe+36>
0x1009e81a <+34>: mvns r0, r1
0x1009e81c <+36>: movs r1, #0
0x1009e81e <+38>: pop {r2, pc}
```There's sure a lot of instructions hereit just be in the fw, not the rom.
but yes the rom is open source too https://github.com/raspberrypi/pico-bootrom
tulip sleet
we could make our own fast paths; would not be hard to bencharmk
cedar veldt
Hi, I asked yesterday in help-with-circuitpython but maybe it's more of a dev question so hope it's OK to bring it here! I'm wondering why the Arduino Nano RP2040 Connect has 50K more RAM available to the user program than the Pico W. (On 8.0.0-beta.3)
manic glacierBOT
tulip sleet
Hi, I asked yesterday in help-with-circuitpython but maybe it's more of a dev qu...
The WiFi coprocessors require different supporting software, since they are completely different, and don't use the same underlying protocol. More work is done on the main-processor side for Pico W than for the NINA-FW.
proven garnet
A heads up the PyPI is in read-only "maintenance" mode right now for scheduled upgrades - I can confirm any releases will fail to upload.
The website also says as much
thorny dove
@proven garnet the 403 error. The reason could be an outage of internet somewhere. I had 403 errors a few days ago clicking on a link on an Adafruit repo towards ReadTheDocs. It turned out that the reason was outage of internet. A few hours later all was normal. Clicking the link then brought me to the docs page. It was on the same day that people in my country (Portugal, EU) reported outage of WhatsApp services.
manic glacierBOT
Taking this pin for example, you might define both D29 and A3 but not any of SCL0, RX0, /CS1, or 6B.
When there's e.g., a single set of pins designated for SPI, UART, or I2C we name them separately. So, for instance, if the use of the pins labeled D0/D1 as TX/RX is well established for this form factor, I might put that alias even though it's not on the silkscreen. Similarly, D...
onyx hinge
@tulip sleet a "power of 2 tile size" optimization won't help with an extremely common case, the terminal text. Its grid is non-power-of-2
I was thinking of tracking the sub-grid unit explicitly, but that got hairy quickly.
tulip sleet
<@329766224093249548> a "power of 2 tile size" optimization won't help with an e...
i meant the only the scale. is the scale 1?
i wasn't thinking about the mod
onyx hinge
the tilegrid scale of the terminal can be non-1 as well, not sure if it's ever not 1 or 2 though
tulip sleet
does divmod get called once or twice?
onyx hinge
oh that's just one example of use of division/modulo, the more complex expression involves tracking both the scale factor and the grid coordinates
uint16_t tile_location = ((local_y / self->tile_height + self->top_left_y) % self->height_in_tiles) * self->width_in_tiles + (local_x / self->tile_width + self->top_left_x) % self->width_in_tiles;
input_pixel.tile = tiles[tile_location];
input_pixel.tile_x = (input_pixel.tile % self->bitmap_width_in_tiles) * self->tile_width + local_x % self->tile_width;
input_pixel.tile_y = (input_pixel.tile / self->bitmap_width_in_tiles) * self->tile_height + local_y % self->tile_height;
it's quite complex, but there are divisions by the scale, and divisions and modulos by the tile width and height
tulip sleet
some discussion this is already optimized in some way: https://stackoverflow.com/questions/5608559/division-and-modulus-using-single-divl-instruction-i386-amd64
onyx hinge
incrementally, it would be good to figure out how to hoist the "y" calculation parts to the outer loop
the divmod_s32s32 function gives the quotient and remainder at once, but as I pasted yesterday there are a lot of function call layers before actually interacting with the fast hardware divider register.
The SIO provides one 8-cycle signed/unsigned divide/modulo module to each of the cores
proven garnet
PyPI is back online!
cedar veldt
The WiFi coprocessors require different supporting software, since they are comp...
Right, it's not surprising that Pico W would be doing more work for WiFi on the main-processor side. But I was still puzzled about why there's such a difference for a program with no WiFi in it. Does it need to do some things on startup?
onyx hinge
There's a lot of RAM reserved for the networking functionality even if it's never used.
cedar veldt
There's a lot of RAM reserved for the networking functionality even if it's neve...
I see, thanks for the info! In my case I do want to be using the WiFi, but I wonder if people will be surprised about Pico programs not working on Pico W.
tulip sleet
@onyx hinge I need to add some forgotten include guards in some files. I see varying opinions about whether #pragma once is safe or not. Files with the same name (of which we have many) are noted as a possible stumbling block, but it seems like the compiler checks much more than that: it would check time, and contents as well. You've used it several places: is there any reason NOT to use it that you know of?
onyx hinge
as far as I know, #pragma once as implemented in gcc is safe and frees you from the need to worry about having unique header guards.
tulip sleet
thanks, it's a lot easier to type, and I think any issues would show up as compilation errors, at least for us
onyx hinge
Identifying the same file on a file system is not a trivial task.[6] Symbolic links and especially hard links may cause the same file to be found under different names in different directories. Compilers may use a heuristic that compares file size, modification time and content.[7] Additionally, #pragma once can do the wrong thing if the same file is intentionally copied into several parts of a project, e.g. when preparing the build. Whereas include guards would still protect from double definitions, #pragma once may or may not treat them as the same file in a compiler-dependent way. These difficulties, together with difficulties related to defining what constitutes the same file in the presence of hard links, networked file systems, etc. has so far prevented the standardization of #pragma once.[citation needed]
I disagree with almost everything in this paragraph on wikipedia 🙂
tulip sleet
exactly, some people in the comments did too; I read that a couple of years ago, and just now
onyx hinge
at $OLD_JOB we had a manually executable test for the include guards matching a defined style. We could code that in pre-commit, but I believe #pragma once is preferable. It's on any plausible compiler and works fine for our needs.
I can vaguely imagine some awful build system that DOES do these symlinking and hardlinking things and could run into trouble, but .. stop hurting yourself like that.
tulip sleet
i just made all the common-hal alarm .h's uniform, in the course of adding guards to a bunch of files that were missing them (it worked by accident up to now)
onyx hinge
should help marginally with compile time too
onyx hinge
huh
tulip sleet
the test case is rather pathological
onyx hinge
something must have gotten better since 2013 when the difference was 21x, and the later discussion in 2018 when it was 5x. Well, I guess the difference is that the guarded version got slower, the pragma once version is about the same. 😭
tulip sleet
it might be lotsa RAM on the test computer or SSD's
onyx hinge
almost every time I invested a performance problem in the gnu toolchain it turned out to be a "boy, this is quadratic, because you've got a linear search, probably through a linked list". Mostly I saw it in binutils. Too bad more sophisticated things like hash tables are so hard to build in C.
I see this is a similar thing
Any use of "#pragma once" sets this flag.
This then leads to a loop over all headers, so n^2 behavior.
and I never managed to fix any of them myself, because I didn't even want to campaign for the time to work on it
tulip sleet
when I was in a compiler class in grad school in 1978 I had to write my own C string library. C++ might be terrible in various ways but at least it has useful libraries
onyx hinge
on a different subject, do we have a writeup of WHY we require a unique vid/pid for each build? I feel like I don't know why, except that it's what adafruit does, and maybe that it's actually important to Arduino.
tulip sleet
i think the two original motivations are for arduino, and for programs like Mu to identify the board. The latter is somewhat moot now that we use the USB interface names
Limor had a reason which I've asked about a couple of times but forget after a while each time
onyx hinge
sounds like we're in a similar boat. I wasn't sure if this was from Limor or from Scott.
onyx hinge
something to do with how they flash & QA the boards?
tulip sleet
that's an interesting point -- does validation during test
onyx hinge
to the extent that it's about Adafruit's practices I don't see why we would make a hard line enforcing it on others. Why not allocate one PID for "generic third party CircuitPython product" and allow that to be used for any non-adafruit board..?
tulip sleet
Limor had a good reason, but I can't remember it now. Worth asking again. It should be from pidcodes, not us.
jaunty juniper
I believe there is an issue with windows caching drivers based on VID/PID, which is why we need to use the cleaner tool sometimes, if each board has the same pid that problem might appear much more often
tulip sleet
yes, that' sounds familiar, thanks
jaunty juniper
by the way I'm not sure how Mu does with ESP32/C3 boards, can we use the Micropython mode or something ?
onyx hinge
Yeah I wouldn't know about Windows problems. But, in fact, CircuitPython boards can change their descriptors all the time since at least 7.x...
jaunty juniper
yep and sometimes I see a board stop being recognized properly because of a change like enabling/disabling part of the USB things, solved with the cleanup tool
tulip sleet
by the way I'm not sure how Mu does with ESP32/C3 boards, can we use the Micropy...
we have said to use Thonny for boards without CIRCUITPY
I don't know that Mu has any support for non-MSC boards
jaunty juniper
(I don't use CP on windows a lot, but I've seen similar things in support questions)
steep musk
Excuse me if this is the wrong group for this .. I created a "Socket Server" using inspiration from the existing HttpServer is that something that you'd be interested in putting into circuit python alongside the HttpServer?
onyx hinge
The USB-IF have given Raspberry Pi permission to sub-license the USB product ID values for its vendor ID (0x2E8A) since they are to be used on a common silicon component which will be used within a customer's product (the RP2040 silicon). For each of the allocations we (Raspberry Pi) will have to apply to the USB-IF to agree this use. This document then contains the full list of product IDs we have approved.
Oh poor rpi, they have to take each request to USB-IF. well, nobody said it should be easy.
I need someone to convince me not to start a website called "nebraska firmware mart" that allocates small integers that, in association with the nubmber 0004, could potentially be used for some unspecified purpose. (Vendor ID 0004 was allocated to Nebraska Furniture Mart, but is listed as abandoned)
onyx hinge
Excuse me if this is the wrong group for this .. I created a "Socket Server" us...
that sounds like something that could belong in the "Community Bundle". https://github.com/adafruit/CircuitPython_Community_Bundle
the library has to follow the usual format of all the adafruit libraries, which is easiest if you create the structure for your github repo using cookiecutter (you can do that, then copy in the files you already wrote)
jaunty juniper
that would be the guide
https://learn.adafruit.com/creating-and-sharing-a-circuitpython-library
steep musk
Thanks both and thanks for all the great work. I've created a WiFi controlled USB hid device over the last couple of days
tulip sleet
> The USB-IF have given Raspberry Pi permission to sub-license the USB product I...
Mrs. Blumkin may have been very far-sighted
jaunty juniper
@tulip sleet I am looking at adding safe mode with the button on the m5stack stamp C3 and because that adds a new safe mode message (the button has a name that is not in an existing string) it also does this in circuitpython.pot, from main. (In addition to adding the string)
-#: py/objnamedtuple.c
+#: py/objexcept.c py/objnamedtuple.c
on "can't set attribute"
tulip sleet
<@329766224093249548> I am looking at adding safe mode with the button on the m...
ah, ok, sorry. I did not see that. So that's you also?
jaunty juniper
no that PR is not me
I think "make translate" was not run when the error message was added to objexcept.c or since, is the reason for that line's change https://github.com/adafruit/circuitpython/pull/7103
idle owl
@onyx hinge @tulip sleet I ran into the same Learn issue with takeover, while Justin was the one looking at my guide page. So, there's an issue filed now for it. 😄
cedar veldt
How do I build Pico W? I'm getting make: *** No rule to make target 'lib/cyw43-driver/src/cyw43_stats.c', needed by 'build-raspberry_pi_pico_w/genhdr/qstr.split'. Stop.
tulip sleet
How do I build Pico W? I'm getting `make: *** No rule to make target 'lib/cyw43-...
did you fetch the submodules?
do make fetch-submodules at the top level
idle owl
@tulip sleet Is there any way to view super-old Adabot library reports on S3? I thought you had a hand in making them only recent ones.
I want something like 1 February 2019.
cedar veldt
do `make fetch-submodules` at the top level
Yes I did
... But maybe I was on the wrong branch when I did it. trying again
jaunty juniper
by the way I make alphas as I progress with my web (and usb) workflow app (apparently around once a week so far), if anyone wants to try it out.
It's very alpha but can edit files, install and update libraries, backup the board's content (useful with web workflow since updating CP with esptool usually erases the flash)
https://github.com/Neradoc/discotool-circuitpython-manager/releases
I make windows and mac builds manually but I think you can run it on linux by running npm install and npm start from the repo dir.
cedar veldt
Fetching the submodules again didn't make any difference.
Ah no, I think that operation didn't complete because of an error in espressif
tulip sleet
<@329766224093249548> Is there any way to view super-old Adabot library reports ...
I deleted those old reports, sorry! I do that by hand every few months. Is the info in the old meeting notes?
idle owl
I deleted those old reports, sorry! I do that by hand every few months. Is the i...
I have no idea. Good to know though. I'll check that at least.
tulip sleet
Ah no, I think that operation didn't complete because of an error in espressif
If this persists, I can take a look. I have not encountered that problem.
cedar veldt
If this persists, I can take a look. I have not encountered that problem.
Something got into the wrong state in git and ports/espressif/esp-idf was dirty. I'll see if it happens again
I'll also see if I can still build espressif sometime 😉
idle owl
@tulip sleet No apparently it wasn't in the notes doc. I checked a few. I'm now not sure where we even stored this info. Remember when we had PyPI download stats for every library? I genuinely don't even remember what we did with that data. I'm suggesting if we can get it going again to add it to the autogenerated library list in the Bundle repo (circuitpython_library_list.md, not drivers.rst), and also include a total for ALL libs in the library section of the daily lib report, since listing every library would be untenable.
thorny dove
@tulip sleet or anyone else who owns an ESP32-S3-BOX, when you have the time for it, can you check on your ESP32-S3-BOX the x,y touch values for the four corners of the display (approximately). See the uploaded 'code_touch_test.py'. In the top of the script it has the values I found on my esp-box. I ask this because my esp-box has the x,y 0,0 in the right-top (instead of left-top).
tulip sleet
<@329766224093249548> No apparently it wasn't in the notes doc. I checked a few....
i kind of remember that, but are pypi downloads that interesting -- they only tell us about blinka use
I think there was some issue about their accuracy? I have a vague memory of that
idle owl
That I don't recall. Maybe? I think it was vaguely interesting, And moreso now because editing environments use them, and so on, so it kind of shows more than Blinka use at this point. I'm going to have Alec look into what's needed to get that data again, and if it's bonkers, forget it. But if it's reasonable, I think it's worth having.
He's also planning to look into the Blinka PyPI download stat, because he thinks that's inaccurate for PyPI, and might be able to get a more accurate and appropriate number for that as well.
tulip sleet
<@329766224093249548> or anyone else who owns an ESP32-S3-BOX, when you have the...
Yes, I see the same thing. I don't know if it's due to the library code (maybe it's inverting things) or due to something else. Also the touch library was written for a particular version of the touch sensor, maybe, and maybe the version on the BOX is different??
jaunty juniper
yup same coordinates
idle owl
Yes!
thorny dove
@tulip sleet & @jaunty juniper thank you
idle owl
@tulip sleet Thank you so much. I felt like the whole situation had ghosted. I swore it was documented....
But I couldn't find anything.
jaunty juniper
the driver is pretty basic, it reads bytes, and splits them into the values with struct, no math done on x or y
idle owl
I remembered the other thing PyPI stats were good for, @tulip sleet. When we do library-wide updates, we know which libs are likely the most popular, and we can avoid starting with those, and start with the bottom of the list instead. That was really nice to have. As it is, we guess.
thorny dove
@jaunty juniper OK
@tulip sleet & Neradoc about the touch controller the ESP32-S3-BOX: It's a bit confusing for me: in github.com/espressif/esp-box/components/i2c_devices/touch_panel folder there are the files ft5x06.c (with I2C address 0x38) and tt2100.c (with I2C address 0x24). I go to check if an I2C.scan() reveals any of them. OK: the i2c.scan found the following addresses: ['0x18', '0x24', '0x40', '0x68']. I had no external devices attached other than that piggy-back 3-led board and an CP2102N via UART. However I have the adafruit_tt21100 module imported and a touch object created using 'tt = adafruit_tt21100.TT21100(i2c).
jaunty juniper
hmmm it stops detecting touches after a few
thorny dove
ahhhh... I thought I had made an error in my program or so. When I push a finger for longer time it ultimately comes with updated x,y values. But this touch part I am using now in a much long script for other things. I just inserted the touch part to test it there in that script.
The script I uploaded above was written solely for a touch test.
In the factory demo the touches work fine.
onyx hinge
curious. I wondered whether it took much time to claim a DMA channel. Keeping the SPI peripheral's DMA channels around from the time it is locked until it is unlocked is slower at doing displayio refreshes...
specific to rp2040
blissful pollen
curious. I wondered whether it took much time to claim a DMA channel. Keeping t...
I was wondering that too but never got a chance to test it
I also think (and may be imagining it) that after the DMA write it, CP pauses until it is done
idle owl
OK, so I'm writing the guide for the PiCowbell Proto (https://www.adafruit.com/product/5200) and there are four ways to assemble it for the Pico(W). The fourth one is as follows: "Solder the PCB directly to the Pico headers - of course this is very compact and inexpensive but you won't be able to remove the PiCowbell." I don't see how that's possible. It's not reversible, the STEMMA QT requires you to have it oriented properly so GPIO4 and GPIO5 are the Pico(W) pins connected to the STEMMA QT connector. So if you try to do that with the board oriented properly, either the components on the Pico(W) get in the way, or the button and STEMMA connector on the PiCowbell get in the way. Am I missing something obvious here? I'll get clarification from Limor on Monday, but I figured I'd ask here in case anyone else saw something I haven't.
Anyway, if anyone has any ideas, please tag me. Half of the guide will be documenting the four options to assemble it. So I have to figure this out eventually. 😄
ember iris
OK, so I'm writing the guide for the PiCowbell Proto (https://www.adafruit.com/p...
If I understand this correctly, it is suggesting soldering the cowbell to a pico which has these male headers soldered like this: https://cdn-shop.adafruit.com/970x728/5584-02.jpg
where the plastic spacer on the male short plug headers offer enough clearance for the stemma and button to be spaced in like that and you can solder the pins directly to the cowbell which make it a permanent design
But I could be thinking about this wrong and not seeing it
idle owl
Ohhhhh, it says "Pico headers", not "Pico". You may be right there.
That makes a lot more sense.
Checking one of my soldered Picos to see if it fits
ember iris
It's my best guess because you're right, a direct solder on the through holes wouldn't make sense here
idle owl
Oof I did a bad job soldering these headers on. One is on an angle, making getting it into the Cowbell difficult. Ugh.
There we go.
OK, yes, that seems reasonable. There's still a little gap between the PiCowbell and the plastic bit on the header, but it's definitely a slimmer sammich than any other option.
Could trim the header pins to slim it further.
ember iris
*nervously stares at my soldering, which often is brute forced into place*
Huh, I *totally* don't know what that's like
idle owl
*nervously stares at my soldering, which often is brute forced into place\*
Huh,...
Hah. I usually do headers while in a breadboard, which means they're usually straight. So... no idea what happened here.
ember iris
I'm getting better at it, but back in college I relied on a few friends who were amazing at soldering. Which might have not been a great option in hindsight, but here we are--with new skills to learn lol
idle owl
I'm getting better at it, but back in college I relied on a few friends who were...
I learned from my bestie who is brilliant at soldering. I haven't worked on my SMT skills, but otherwise my soldering is really good. Learning from him helped immensely.
onyx hinge
I also think (and may be imagining it) that after the DMA write it, CP pauses u...
yes, it does wait for the DMA to complete, which is the usual requirement for how we do spi/i2c/etc transfers. I experimented with whether NOT doing run_background_tasks (but just busy looping) helped anything but it .. didn't.
it's a bummer I don't know how to do function-level profiling on microcontrollers, at this point I think I'm probably guessing wrong about what's taking a long time
cobalt grail
Can someone clarify what BOARD_USER_SAFE_MODE_ACTION is supposed to explain?
cobalt grail
It would also be helpful if someone could tell me how to get rid of the stupid "wat" icon that automatically got added to my post?
jaunty juniper
so you know why it happened
example:
You are in safe mode because:
You requested starting safe mode by pressing G3 button at start up.
To exit, please reset the board without pressing G3 button at start up.
(that's a build I'm looking at for the M5Stack stamp C3)
(I added a "the" since)
on many boards it would be "boot button"
the CPX says: pressing both buttons at start up
cobalt grail
So it is an explanation but only half of the sentence? And no. I have no idea why it happens.
jaunty juniper
So it is an explanation but only half of the sentence? And no. I have no idea wh...
I meant the string is there so you know why it happened 😛
cobalt grail
And yes, it was M5Stack only.
jaunty juniper
basically it's 2 sentences that are stuck at compile time:
You requested starting safe mode by + BOARD_USER_SAFE_MODE_ACTION
To exit, please reset the board without + BOARD_USER_SAFE_MODE_ACTION
it is already on many boards, you could lookup how they do it
the "both buttons version" and "boot button" version etc.
does that work ?
https://hosted.weblate.org/browse/circuitpython/main/en/?offset=1&q=pressing&sort_by=-priority%2Cposition&checksum=
https://hosted.weblate.org/translate/circuitpython/main/en/?checksum=31cefe380cd4850d&q=safe+mode&sort_by=-priority%2Cposition
https://hosted.weblate.org/translate/circuitpython/main/en/?checksum=c1b5d85d73cafbab&q=without&sort_by=-priority%2Cposition
cobalt grail
Ok. For anyone "listening" - You can't use half of an english sentence and add another half of a sentence and think it may make sense if string concatenating the two.
jaunty juniper
yeah
cobalt grail
That is not how other languages work.
jaunty juniper
we should rework it into a separate sentence
cobalt grail
So give me a full sentence to translate. Don't try to be smart. It does not work.
idle owl
It would also be helpful if someone could tell me how to get rid of the stupid "...
Only you can get rid of it. Edit the post and delete it, I believe. I can delete your entire post, but that's not the right thing here. I can't edit it.
cobalt grail
Thanks for the link. I fixed the symantics by adding 2 words. Just wish I knew where to add those words in the sentence prefix, as the are the same for all.
cobalt grail
Only you can get rid of it. Edit the post and delete it, I believe. I can delete...
I am more like; Why did Discord even ADD IT?
I tried to edit my post uyt nope. Very odd!
idle owl
I am more like; Why did Discord even ADD IT?
I have no idea, I think there's a way to click on them, and I've seen it pop up and have had to avoid accidentally adding those things. So they're so easy to add, but apparently really annoying to remove.
jaunty juniper
it's like it tries to push you to use it with popups so you have to intentionally cancel it or something
is there a X next to it when you hover like link previews ?
cobalt grail
Nope. No way to delete it. Very strange.
It does state that is the "What" decal so.
Either way, I will spend exactly 0 seconds on that going forward.
jaunty juniper
🤷
jaunty juniper
how about:
You are in safe mode because:
Both buttons were pressed during start up.
To exit, please reset the board without requesting safe mode.
LGTM! + Its smaller 👍
jaunty juniper
that is a bonus
analog bridge
If you are at it, maybe replace:
- G3 button was pressed, BOOT button was pressed
+ %q button was pressed, MP_QSTR_BOOT
not sure how that fits, I'll open the PR an we can discuss it there
lavish saffron
I'm adding typing to the Adafruit_CircuitPython_Wiznet5k library.
When I try to build the documents with sphinx I get the following exception:
Warning, treated as error:
autodoc: failed to import module 'adafruit_wiznet5k' from module 'adafruit_wiznet5k'; the following exception was raised:
Traceback (most recent call last):
...
File "/Users/martinstephens/Documents/Coding/Repos/Adafruit_CircuitPython_Wiznet5k/adafruit_wiznet5k/adafruit_wiznet5k.py", line 43, in <module>
import adafruit_wiznet5k.adafruit_wiznet5k_dhcp as dhcp
File "/Users/martinstephens/Documents/Coding/Repos/Adafruit_CircuitPython_Wiznet5k/adafruit_wiznet5k/adafruit_wiznet5k_dhcp.py", line 28, in <module>
import adafruit_wiznet5k.adafruit_wiznet5k_socket as socket
File "/Users/martinstephens/Documents/Coding/Repos/Adafruit_CircuitPython_Wiznet5k/adafruit_wiznet5k/adafruit_wiznet5k_socket.py", line 28, in <module>
_the_interface: Union[wiznet5k.adafruit_wiznet5k.WIZNET5K, None] = None
AttributeError: module 'adafruit_wiznet5k' has no attribute 'adafruit_wiznet5k'
The code in question is below and PyCharm is not flagging this reference as a problem. I'm stumped.
try:
from typing import Any, Optional, Tuple, List, Union
except ImportError:
pass
import gc
import time
from micropython import const
import adafruit_wiznet5k as wiznet5k
# pylint: disable=invalid-name
_the_interface: Union[wiznet5k.adafruit_wiznet5k.WIZNET5K, None] = None
Here's the directory tree, which makes me think that the reference is good:
├── adafruit_wiznet5k
│ ├── __init__.py
│ ├── adafruit_wiznet5k.py -> WIZNET5K is a class
│ ├── adafruit_wiznet5k_dhcp.py
│ ├── adafruit_wiznet5k_dns.py
│ ├── adafruit_wiznet5k_ntp.py
│ ├── adafruit_wiznet5k_socket.py
│ └── adafruit_wiznet5k_wsgiserver.py
Any help greatly appreciated as always!
violet thistle
Hey. asyncio.run() is not working in V8.0, is this intentional or an oversight?
stuck elbow
it's a bug, there is a fix already
proven garnet
So it looks like the weekly report only reports downloads from piwheels.org, which doesn't include good ol' PyPI. We actually (and unsurprisingly) get way more downloads from PyPI, about 25322 in the last week (10903 was the reported piwheels for reference). Working on adding that back into adabot and the weekly meeting since I got Google BigQuery to work with pulling that data, which is how I'm getting that number.
Not sure I'll have it in time for next meeting but effectively if you add both of those, for example, the actual number of times Blinka is downloaded shoots from 10903 to 36225, over triple the reported amount!
manic glacierBOT
It happens mainly in Mu:
In PuTTY the CPY V8 statusbar text is seen instead of 'PuTTY' in the caption of the PuTTY app window only
but not in the print output:
: global tt_touched cnt = 0 ret = True tt_cnt = 0 while tt.touched: ts = tt.touches if isinstance(ts, list) and len(ts) > 0: tt_cnt +=1 cnt += 1 if cnt >= 100: ret = False break print(f"clr_touch(): touches count= {tt_cnt}") if tt_cnt > 0: tt_touched = False The global tt_touched flag is set when there was a touch event. This is done in function ck_touch(). Then, in function loop() there is this call to clr_touch() when the tt_touched flag is set: if tt_touched: clr_touch() REPL output after I touched the display: ```
ck_touched(): screen touched at: [{'y': 106, 'x': 144, 'id': 0, 'pressure': 34}]
menu_touch(): Touched: id= 0, at x,y = 144,106. Pressure= 34
clr_touch(): touches count= 3```Conclusion: the touch module, for reason unknown to me, fills the touch event 'buffer' with three occurrances/events after one touch. I also saw one time (or maybe few more times) that the 'touch event buffer' already had touch event data at start of the script. Perhaps I touched the display before (or touching it with one or more fingers to fix the device onto the table while pushing the reset button on the side). Anyway I implemented a call to clr_touch() at the start of the script and in the loop(), when tt_touched is set I do a call to clr_touch() to have a 'clean touch slate'. This solution works in this script as I like it to be. After a touch I now only get one call to a function 'menu_touch(ts)' which displays the touch event attributes: id, x, y and pressure'. The ultimate goal is to create a menu-system using touch events/data.
thorny dove
@tulip sleet the touch attributes displayed
tulip sleet
<@329766224093249548> about the touch readings on the ESP32-S3-Box I wrote this ...
If you see changes you would like to make to the library, issues or pull requests are welcome.
thorny dove
OK, I did not (recently) see the contents of that library. I go to do that.
I realize myself that I have to keep in mind that the code for the touch events (as a lot of other events), 'outside of my CPY script' are always running in the background. So, if I touched the display before starting my script, the code of the device already has accounted the fact that the display been touched.
tulip sleet
@onyx hinge I wrote the release notes for 8.0.0-beta.4. If you feel like reviewing #7154, I'll make a release today. I tested it and a user tested it.
onyx hinge
it touches 47 files, eek!
tulip sleet
many are just renaming, or the same change multiple places
onyx hinge
create -> record is a renaming?
tulip sleet
those routines used to actually allocate a new object. Now they record the info in a static singleton, hence the name change.
also "wakeup" -> "wake" for consistency.
the singleton is a union to hold the largest possible object. (even the largest is pretty small)
tulip sleet
I started to add even more flags that handled the lifetime of the alarm info, but it was a mess, and varied per port. This is simpler.
onyx hinge
simplification is good
jaunty juniper
OK, I did not (recently) see the contents of that library. I go to do that.
you might want to find the datasheet for the sensor, because the library is written for another sensor, isn't it ?
orchid basinBOT
vague thicket
Hello, can anyone point me to something which would help me fix the issue with CI? I have PR which is failing (https://github.com/adafruit/circuitpython/actions/runs/3356536863/jobs/5561676282), but I do not know what to change to make it finish. I understand that there is problem with translation and I sync with Adafruit CPY repo, but that does not help. I do not have these files in my PR so I´m bit confused and I did not find any topic in history which would help me. I´m new to the CI so sorry for probably stupid question. Many thanks for any help.
tulip sleet
Hello, can anyone point me to something which would help me fix the issue with C...
run make translate at the top level of your repo and then commit that change
do you mean that you merged in the latest from adafruit? You may still need to do make translate
vague thicket
do you mean that you merged in the latest from `adafruit`? You may still need to...
thanks @tulip sleet , that helped resolve my issue 🙂 now I´m waiting if all checks will be ok..
manic glacierBOT
manic glacierBOT
lavish saffron
Hi, I've been adding typing to the Adafruit-CircuitPython-Wiznet5k package. I have hit a roadblock with adding typing to the WIZNET5K class. The class is used as an argument for classes in other modules (e.g. DNS, DHCP) that are in turn imported into the wiznet5k module. Thus, when I import the WIZNET5K class to add as a type I create cylic imports. This prevents import adafruit_wiznet5k.adafruit_wiznet5k from working and makes pylint grumpy. I've tried some suggestions from the Interweb and I've grep'd the entire CircuitPython bundle for cyclic-import. I'm stumped.
Any suggestions will be gratefully received!
jaunty juniper
I believe that's when you need to start with from __future__ import annotations as the first import
not sure though
onyx hinge
if it's pylint, you can add the right magic comment to tell it not to produce the message in the first place. Here's one example: https://github.com/adafruit/Adafruit_Blinka/blob/ef2657eaa1caf3262c6bbe1f2df5b6ed08f72162/src/adafruit_blinka/microcontroller/generic_micropython/__init__.py#L21 ```py
def __repr__(self):
# pylint: disable=import-outside-toplevel, cyclic-import
import board
import microcontroller.pin
for key in dir(board):
if getattr(board, key) is self:
return "board.{}".format(key)
# pylint: enable=import-outside-toplevel, cyclic-import
``` though I don't know if there's a reason that the one-line version import microcontroller.pin # pylint: disable=cyclic-import wasn't used, I'd try that first
lavish saffron
not sure though
That's spot on, thanks!
lavish saffron
I'm getting this error while building the docs with sphinx:
...adafruit_wiznet5k.py:docstring of adafruit_wiznet5k.adafruit_wiznet5k:6:Field list ends without a blank line; unexpected unindent.
How do I find the location of the error within the file (1000 lines of code)? I've added -v -v to the command but it's still obtuse! Thx
lavish saffron
I'm getting this error while building the docs with sphinx:
`...adafruit_wiznet...
Found it. `\r\n in a doc-string. Grrrrr 
manic glacierBOT
proven garnet
Found it. `\r\n in a doc-string. Grrrrr <:ruby:342384086205333505>
I think there's a precommit hook for this that I want to add
tulip sleet
@proven garnet @idle owl @jaunty juniper circup is throwing an error getting a library from the latest bundle. My circupt version is up to date. I haven't figured it out yet but need to do some other things. If you have time, could you take a look
jaunty juniper
what error ?
tulip sleet
halbert@cod:~$ circup install adafruit_lc709203f
Found device at /media/halbert/CIRCUITPY, running CircuitPython 8.0.0-beta.3-20-g2bba6a68e.
Downloading latest version for adafruit/Adafruit_CircuitPython_Bundle.
[####################################] 100%
[####################################] 100%
[####################################] 100%
OK
Searching for dependencies for: ['adafruit_lc709203f']
WARNING:
typing-extensions is not a known CircuitPython library.
WARNING:
typing-extensions is not a known CircuitPython library.
Traceback (most recent call last):
File "/home/halbert/.local/bin/circup", line 8, in <module>
sys.exit(main())
File "/home/halbert/.local/lib/python3.10/site-packages/click/core.py", line 1128, in __call__
return self.main(*args, **kwargs)
File "/home/halbert/.local/lib/python3.10/site-packages/click/core.py", line 1053, in main
rv = self.invoke(ctx)
File "/home/halbert/.local/lib/python3.10/site-packages/click/core.py", line 1659, in invoke
return _process_result(sub_ctx.command.invoke(sub_ctx))
File "/home/halbert/.local/lib/python3.10/site-packages/click/core.py", line 1395, in invoke
return ctx.invoke(self.callback, **ctx.params)
File "/home/halbert/.local/lib/python3.10/site-packages/click/core.py", line 754, in invoke
return __callback(*args, **kwargs)
File "/home/halbert/.local/lib/python3.10/site-packages/click/decorators.py", line 26, in new_func
return f(get_current_context(), *args, **kwargs)
File "/home/halbert/.local/lib/python3.10/site-packages/circup/__init__.py", line 1310, in install
device_modules = get_device_versions(ctx.obj["DEVICE_PATH"])
File "/home/halbert/.local/lib/python3.10/site-packages/circup/__init__.py", line 903, in get_device_versions
return get_modules(os.path.join(device_path, "lib"))
File "/home/halbert/.local/lib/python3.10/site-packages/circup/__init__.py", line 954, in get_modules
metadata = extract_metadata(source)
File "/home/halbert/.local/lib/python3.10/site-packages/circup/__init__.py", line 536, in extract_metadata
if loc > -1:
UnboundLocalError: local variable 'loc' referenced before assignment
ok that's what happens when a mpy file is invalid
tulip sleet
https://github.com/adafruit/circup/issues/146 and maybe others
yes, so something odd about the .mpy's in the bundle? The previous problem was a badly assembled bundle
jaunty juniper
can you try with an empty lib directory ?
turbid radish
@lone axle - newsletter draft is ready to crib off of for notes
onyx hinge
proven garnet
Interesting I can take a look this evening perhaps if it isn't resolved by then
lone axle
Thank you
proven garnet
Update to my Piwheels comment - it was 11337 downloads from Piwheels in the last month, but 31466 from PyPI in the last week
jaunty juniper
yes, so something odd about the .mpy's in the bundle? The previous problem was a...
I can't reproduce, I think it's choking on a file in your lib directory, can you try a circup list with this branch ?
https://github.com/Neradoc/circup/tree/wrong-mpy-file-response
proven garnet
So it's more like 12 times the number of downloads 🤯
tulip sleet
I can't reproduce, I think it's choking on a file in your lib directory, can you...
Version 0.9.10.dev34+g49860ce of circup is outdated. Version 1.1.3 was released Friday September 16, 2022.
Found device at /media/halbert/CIRCUITPY, running CircuitPython 8.0.0-beta.3-20-g2bba6a68e.
The following modules are out of date or probably need an update.
Major Updates may include breaking changes. Review before updating.
MPY Format changes from Circuitpython 6 to 7 require an update.
Module Version Latest Update Reason
-------------------- ----------- ------- --------------
adafruit_lc709203f 2.2.2 2.2.8 Minor Version
adafruit_scd4x 1.3.3 1.3.7 Minor Version
adafruit_bus_device Bad format 5.2.3 Major Version
a bit weird because I've been using circup on 8.0.0 builds with no issues
(I think!)
proven garnet
It's a day for spooky tricks!
jaunty juniper
is something abnormal with the adafruit_bus_device file then ?
tulip sleet
I wiped lib/ and was able to isntall the library. So I guess there were old versions there
perhaps it should just erase and reinstall if it finds a "bad" library
my local problem is solved, thanks!
proven garnet
Library ghosts 
jaunty juniper
I wanted to add a unit test for bad MPY, and maybe there was something else, I don't remember why I didn't open that PR, I'll do it now and we'll see
lone axle
<@&356864093652516868> the weekly meeting is pumpkin spice flavored, and will occur at 2pm eastern today, a bit over an hour from now. If you'd like to participate join us here on discord and put any hugs or status updates you have in the notes doc. See you there. 🎃
onyx hinge
bonk
onyx hinge
Newsletter looks great based on the sneak preview, Anne!
turbid radish
Thanks!
idle owl
Thanks! 😊
proven garnet
Yes! We can take some extra days to get them merged!
idle owl
🎉
turbid radish
👏
midnight ember
I love all Jeplers keyboard projects. So many keyboards, so little time.
As if desoldering feathers wasn't bad enough now the pico has like 40 pins, that's a lot of desoldering.
onyx hinge
can you make a fritzing diagram of that? I am not understanding your explanation.
onyx hinge
I've been thinking of getting a usb c / displayport hub which I think is supposed to be possible with my laptop, but I'm not confident what to look for / what to get...
midnight ember
i didn't even catch that melissa, that's funny
lone axle
On my end she started metnioning USB / audio all working, then silence 😅
midnight ember
oof he went robot
gilded cradle
You're cutting out @lone axle
idle owl
Sound good now.
midnight ember
good now
gilded cradle
👍
idle owl
@lone axle You disappeared again.
midnight ember
back
tulip sleet
start at top of tekktrik
idle owl
That's entirely fair.
idle owl
never heard of that
I only know of it because of starting on this PyPI project a few years ago.
proven garnet
Yeah, this was one of the challenges of the pastebin library - I had something like a unified API and all fairly similar in use, but they're effectively different things under the hood.
idle owl
@tulip sleet Might have something to add?
proven garnet
Oh I did something similar to this for my personal project, where PRs triggered things in another repo. I can look into it.
Text only
proven garnet
Yeah that's right
midnight ember
github trolling?
onyx hinge
I think they probably mean well
I just don't have the experience that indicates that they're right about it
midnight ember
maybe someone from hacktoberfest trying to get in some PR's somehow?
proven garnet
Sounds good, I can close the suite of issues with what's been said here, I think this makes sense
onyx hinge
"I am not a robot. I am doing a project to help developers with API compatibility checks. I hope to submit a PR or communicate with you, if possible." -- https://github.com/PyDeps
gilded cradle
Thanks
onyx hinge
👋 have a good week all!
austere acorn
I missed the meeting, darn daylight saving changes 🙄 lol
Happy Halloween everyone!
midnight ember
Happy Halloween!
onyx hinge
oh oh! next week the US time changes
idle owl
I missed the meeting, darn daylight saving changes 🙄 lol
Ours is next weekend. Yeah.
midnight ember
Don't forget Help Desk is tonight... though I'm on candy duty.
About what times is Help Desk going to go for? I want to be there but I get a barrage of kids every Halloween in the suburbs so I don't want to have to go in and out.
random junco
About what times is Help Desk going to go for? I want to be there but I get a ba...
7-10pm EDT
idle owl
@tulip sleet I disconnected you from chat so you don't end up with surprises.
tulip sleet
<@329766224093249548> I disconnected you from chat so you don't end up with surp...
thanks, I would do the same for others; writing up the Ventura issue
idle owl
thanks, I would do the same for others; writing up the Ventura issue
Please include a clear explanation of how to complain to Apple effectively. If it makes sense. (I guess I meant to mention this for the blog post which you already did, I think.)
If Apple won't touch it because we didn't complain loud enough, let's incite more complaints.
tulip sleet
I said to use the Feedback Assistant tool. I was just about to encourage it more. pt and I are corresponding with the RPi folks.
idle owl
Excellent, thanks.
jaunty juniper
it's about UF2, is there anyone at Microsoft that knows someone at Apple ? 😏
lone axle
Here is the notes document for next Monday’s CircuitPython Weekly meeting. It is at the normal time of 11am Pacific / 2pm Eastern here on Discord. Everyone is encouraged to attend! Please add your hug reports and status updates even if you’ll be attending the meeting - it’s super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and we’ll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/1BtwVBaXLXcFHM52SGlMX0-7K0m3r_e-T6l6MLfux5KI/edit?usp=sharing
Reminder that daylight savings time ends for US timezones on 11/6. Clocks will be adjusted back by 1 hour the day prior to the weekly meeting.
edgy edge
has anyone ever looked into implementing interrupt handling for an IMU like LSM6DS? The motivation behind this is to enable orientation change and significant movement interrupts to be generated while in very low power state to extend battery life.
lone axle
@idle owl okay, post meeting stuff is wrapped up. I am available now or any time.
idle owl
<@330227457296957440> okay, post meeting stuff is wrapped up. I am available now...
Let's vid!
proven garnet
Closed those issues regarding pinning dependencies, thanks y'all!
idle owl
Closed those issues regarding pinning dependencies, thanks y'all!
Cheers!
@onyx hinge Good call on those issues. I hadn't looked into it, so I didn't respond, but you had it nailed down.
quasi veldt
@meager fog https://blog.adafruit.com/2022/09/12/circuitpython-wifi-support-on-the-raspberry-pi-pico-w-part-deux/ what IDE setup do you have in this first video? Is that Atom + some custom terminal plugin?
idle owl
<@340718517408104450> https://blog.adafruit.com/2022/09/12/circuitpython-wifi-su...
Please do not tag Ladyada. It's entirely inappropriate, and you won't receive an answer. Post your question and give our community a chance to assist. As well, you'd be better off posting to #help-with-circuitpython for questions like this. We discourage crossposting, but in this case, I'm suggesting it.
quasi veldt
Please do not tag Ladyada. It's entirely inappropriate, and you won't receive an...
Happy to put it there.
If you don't mind me asking, I'm not sure why it's "entirely inappropriate". Can you please explain?
Here's what was going thru my mind:
- she made the video I'm linking to (or at least I think it's her voice). She has the prerogative to ignore my message, and anyone else can respond as well.
- And it's related to testing and coding a new library support in circuitpython which I'm happy to help code (I've not written any circuitpython library code yet, but you can see I've participated in design discussions before).
- I want to get involved with committing code (hence I'd like my IDE setup to match)
- Obviously she's the CEO of a company (so am I) and there are boatloads of other things to focus on. I take no offense if she doesn't respond.
What am I missing in my internal logic?
I appreciate your constructive criticism.
No passive aggressive or anything from my side - just looking to learn 🙂
Thanks!!
idle owl
Happy to put it there.
If you don't mind me asking, I'm not sure why it's "ent...
We discourage tagging anyone with whom you were not already in a conversation, or did not have previous permission to do so. There are a couple of reasons for this. One: This community will likely be able to help you sooner than the person you tagged. Two: Tagging someone can imply that you did have that permission or a previous conversation going, and other folks may be less likely to respond. The second one is probably the more important one. Even if the question pertains to an individual, it is still better not to tag them without permission for both of those reasons. It's not listed explicitly in the Code of Conduct, but perhaps it should be.
I understand where you're coming from, as to why you tagged her on that message, but as admins and moderators, we encourage folks to avoid that practice, and instead, simply post the question and let the community help you. Specifically regarding Ladyada, she doesn't really spend time on Discord at all, and does not have notifications turned on. This means if the community assumed you already had help, you likely would not have gotten an answer, which I'm almost certain someone in the community can provide. Due to her commitments and obligations elsewhere, we are more likely to ask folks not to tag her specifically. Even with her not getting the notifications, it has the possibility to get through somehow, and we are protective of her time.
quasi veldt
We discourage tagging anyone with whom you were not already in a conversation, o...
I appreciate the well-reasoned answer
idle owl
I appreciate the well-reasoned answer
You're welcome. Thank you for being receptive.
onyx hinge
oof that is a heavy dependency, rust is also a swiftly moving target at this time. I wonder if we can select a different minifier.
ruby atlas
everything's a moving target 🙂
ornate breach
everything's a moving target 🙂
Except for the observer in the vehicle 🙂
Well, everything in the vehicle is not moving to the observer
onyx hinge
hi rose.
ruby atlas
hi jeff.
ancient hemlock
Is there a quickstart/guide of some kind to adding functionality to CircuitPython? I'm not especially well versed in C/C++ but I like to learn how things work.
stuck elbow
Is there a quickstart/guide of some kind to adding functionality to CircuitPytho...
the deep dive videos with Scott are very useful for that
there is also the build guide, which is something you will need to do anyways https://learn.adafruit.com/building-circuitpython/
otherwise, I would say look at the pull requests that add a module
ancient hemlock
the deep dive videos with Scott are very useful for that
Thanks, appreciate the info. Could you share a link to the YT channel?
ancient hemlock
https://www.youtube.com/playlist?list=PLjF7R1fz_OOXBHlu9msoXq2jQN4JpCk8A
Awesome! Thanks for giving me a place to start.
stuck elbow
scroll down to the ones with Scott, those are the ones about the internals
the later ones with Tim are also interesting, but they are more about the python libraries
ruby atlas
Who do I have to beg for hacktoberfest-accepted labels 🙂
random junco
Who do I have to beg for hacktoberfest-accepted labels 🙂
Tekktrik I believe
ruby atlas
🙂
proven garnet
Hallo
If the PRs were merged in a repository with the hacktoberfest label you're probably fine. Though not sure what happens now that it's November
jaunty juniper
#help-with-circuitpython would be a better place to ask that (and I will answer to you there)
proven garnet
I meant to ask - I keep seeing PRs to ESP_ATControl, which has a depreciation notice in the README (or a no-longer-supported one at least). Do we want to archive that repo?
ruby atlas
@proven garnet they're for here 🙂
proven garnet
I don't think I understand what you mean, do you mean the recently GitHub bot posted PRs need the label?
ruby atlas
https://github.com/adafruit/circuitpython/pull/7161 🙂 even if it doesn't end up being merged in that form, or merged at all. The purpose was to start acting on some of my complaints about going from 0 to working circuitpython build env without missing things 🙂
proven garnet
Out of curiosity, what does your Hacktoberfest account list for that PR? Is it pending?
proven garnet
I can add the label, curious what happens - we definitely don't want it as not-accepted
ruby atlas
both of the late-night ones were. we're still in the 7 day window where the PRs can be labelled.