#circuitpython-dev
manic glacierBOT
robust quartz
Hi I fork the circuitpython, add a new board and generate successfully, right now I have problem on this step , there are many "board" list below the "matrix" line, I don't know which one is, could anyone help please
lunar gull
I have updated the MagTag Air Quality station code to use the newest firmware for the MagTag with proper sleep and such:
https://github.com/NascentMaker/AirQualityStation
tulip sleet
Hi I fork the circuitpython, add a new board and generate successfully, r...
This is no longer necessary; it figures out the new boards dynamically based on the boards/ directories
I have deleted that section in the guide.
robust quartz
This is no longer necessary; it figures out the new boards dynamically based on ...
make sense thank you @tulip sleet
manic glacierBOT
tulip sleet
@lone axle thanks for being around, all this isn't urgent, but may as well get the merges built
tulip sleet
proven garnet
What a gift!
timber mango
@tulip sleet i gave circuit python a try. and omg it's so much better than programming in
tulip sleet
<@!329766224093249548> i gave circuit python a try. and omg it's so much better ...
in ? π
tulip sleet
Thanks for compliment! It is our goal to make it a lot easier.
mental nexus
As the days start to get longer, itβs a good time of year to count blessings, just wanted to express my gratitude to Adafruit and the CircuitPython development team and extended CP library developer/sample code community. While software and hardware sometimes are solitary activities, itβs always nice to connect with kind and welcoming folks willing to lend a hand. That makes each of us better and helps us all grow. Many thanks yβall!
tulip sleet
You're welcome, and thank you for being a great member of the community!
manic glacierBOT
manic glacierBOT
teal thorn
What is the cleanest CircuitPython port to use as a template for a new port?
tulip sleet
What is the cleanest CircuitPython port to use as a template for a new port?
which chip or chip family to do you want to port to?
teal thorn
NXP mcux-sdk. I know there is a mimxrt port. I'll use some of that, but I want to follow the best structure for supporting several families.
tulip sleet
NXP mcux-sdk. I know there is a mimxrt port. I'll use some of that, but I want ...
All the current examples are compromises in one way or another when it comes to incorporating a HAL layer like the sdk you pointed to:
atmel-samd: we had to use a snapshot of a HAL dump from https://start.atmel.com for ASF4. ASF3 was easier: we just used the sdk zip files.
STM: some HAL stuff is just copied
nrf: we use the nrfx low-level HAL, but not the nrf SDK. We had to patch nrfx a bit, and we have our own fork. In addition we use the SoftDevice
espressif: we use their SDK, which requires special build setup
RP2040 ("raspberrypi"): perhaps the simplest: we are using the pico-sdk, but there is some hackery to get the compile to work
In general most of the low-level modules are similar and atmel-samd, nrf, and rp2040 are all good examples. STM is a bit worse because the chips vary a lot, with minor but annoying differences between two similar chips
It also depends on where the firmware is stored. i.MX is tricky because it's all external flash. Same for RP2040 but a little easier. Which NXP chips are you thinking about?
this MCUXpresso SDK Builder looks like start.atmel.com a bit: you customize and download it. That can make it difficult to merge in fixes from upstream unless you set up automation to do that consistently. Ideally create a new separate repo with the SDK in it that you can update as needed. Then we included it as a submodule.
teal thorn
NXP posted the SDK to GitHub so we can submodule it. The API's are mostly universal, but there are some build differences, especially for parts with external flash. We've already got three families in TinyUF2, so I'll probably start with those.
tulip sleet
NXP posted the SDK to GitHub so we can submodule it. The API's are mostly univer...
Would it make sense to revise the i.MX port to use MCUExpresso?
Which chips do you have in mind for this? i.MX, or simpler ones?
orchid basinBOT
It seems like the CI actions for this stalled for a while and eventually failed during the deploy task. Here is the output on the actions page:
I'm not sure what needs to be done to correct it.
robust quartz
Hi I build the program is working on my dev, and I check the error, there is the UNIX board problem and I did not do any change to that board, please give some advice. thank you.
viscid pine
Hi I build the program is working on my dev, and I check the error, there is t...
I think it's a github error, the submodules failed to checkout
so nothing to do with your changes
viscid pine
i believe it doesn't matter, the devs can merge it anyway if they see it's an unrelated error
robust quartz
meke sense, thank you @viscid pine
wait for the good news
warm stump
4 different ports of python:
Top Left: avr8 LUFA USB-to-Serial with micropython embedded
Bottom Left: avr8 LUFA USB-to-Serial (no python, fits in 16KB)
Middle: Micropython Debian x64 Unix Minimal Port
Right: CircuitPython 7.0.x Adafruit Macropad RP2040 Port
warm stump
quite unbelievably 7.0.x worked on the macropad, but I did have to do the following:
sudo apt-get install git build-essential make gettext python3 python3-pip python3-jinja2
sudo pip3 install --upgrade pip
sudo pip3 install huffman
sudo git clone https://github.com/adafruit/circuitpython cp7
cd cp7
sudo git checkout 7.0.x
sudo git submodule init
sudo git submodule update --recursive
sudo make submodules BOARD=adafruit_macropad_rp2040
sudo make BOARD=adafruit_macropad_rp2040
cd build-adafruit_macropad_rp2040
sudo cp -rf ./* /mnt/d/cpy-debian-macropad/ #from Debian WSL
onyx hinge
seeing git pull acting very slow. https://www.githubstatus.com/ only says that actions is running in degraded status.
thorny jay
Oh, it is Monday again already. I have done nothing meaningful or really useful this week. I tested 7.1.0RC on Matrix portal and used a learn guide to display CO2 level during Xmas evening. I had one fight with circup and could not solve it, I guess something is broken, because of 7.1 so I had to update the library manually, but everything worked find. One thing I also noticed is that candle in the room do increase the CO2 level a lot, and this is not a good indicator that the room need ventilation... so it did not work as expected.
I don't think I am going to fill the doc. Hug report to all the contributor of CP core, library, documentation, learn guide.
tulip sleet
Oh, it is Monday again already. I have done nothing meaningful or really useful ...
Update circup to 1.0.3 and it should fix the problem. Please file an issue if you are having trouble. I made circup notice that it is out of date and report that (but of course not on the older version).
thorny jay
Update circup to 1.0.3 and it should fix the problem. Please file an issue if yo...
Thanks. You said something about circup in last meeting, so I figured you were already on top of it. I was just in a hurry to get the CO2 meter working.
tulip sleet
Reminder that we are having a regular CircuitPython Weekly Meeting today in #circuitpython-dev and the CircuitPython voice channel. We expect light attendance due to the holidays. Don't feel obligated. It is at the normal time of 11am Pacific / 2pm Eastern here on Discord. Here is the notes document for Mondayβs CircuitPython Weekly meeting. Please add your hug reports and status updates even if youβll be attending the meeting - itβs super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and weβll read them off during the meeting.<@&356864093652516868>
https://docs.google.com/document/d/1NBK5WWMLkzOm3BYvKe3zdyLp1Z5S5DQfyDGdEudic5w/edit?usp=sharing
onyx hinge
I have a phone call to make, I may be a tiny bit late joining the meeting.
idle owl
I have a phone call to make, I may be a tiny bit late joining the meeting.
No worries.
@tulip sleet I made the new folks in the State of section bold for you.
onyx hinge
hmmmm google docs autocompletes "January 3," with "2021". π
river quest
hi hi folks, thanks @tulip sleet for kicking it off and DJing today! (pt here - in the @river quest account)
timber mango
can i get that role please.
onyx hinge
The calendar is up to date for the 2022 meeting schedule, and can be used online or in a variety of calendar apps included google
idle owl
can i get that role please.
Done!
timber mango
thanks!
meager fog
hiiiiiii
river quest
hi hi
turbid radish
Giving Anne a break
tulip sleet
idle owl
π I did!
tulip sleet
onyx hinge
insert pastry, receive pcb ?
meager fog
hihi
idle owl
ππ» Later!
river quest
cheers!
blissful pollen
Great to have you stop by
idle owl
Oh foamyguy is here!
idle owl
@lone axle I'll leave it to you!
blissful pollen
is that my PR? If so I should really get to updating it asap. Probably not much work
onyx hinge
- Merged pull requests:
- https://github.com/adafruit/Adafruit_CircuitPython_VL53L0X/pull/21 (Days open: 525)
πππππππππππππππππ&ΒΌ months
- https://github.com/adafruit/Adafruit_CircuitPython_VL53L0X/pull/21 (Days open: 525)
idle owl
is that my PR? If so I should really get to updating it asap. Probably not much ...
No idea! You can check https://circuitpython.org/contributing - the days open are listed.
Really if it's being worked on, it's fine. I'm more concerned about the ones that are stagnant, or missed, etc.
onyx hinge
Thanks & see you next year!
errant grail
Thanks!
solar whale
π
onyx hinge
interesting, if you with NeoPixel(...) as n: ..., the pixels all get blanked at the end of the with-statement. Same if .deinit() is called, but not if you run to the end of code.py.
blissful pollen
Is .deinit() not ran at the end of code.py then? Or (just a guess) does the VM clear all the objects but then deinit pin steps (like clearing the pixels) isn't ran?
onyx hinge
seems like.
tulip sleet
Here is the notes document for Mondayβs 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 youβll be attending the meeting - itβs super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and weβll read them off during the meeting. Hope to see you there! <@&356864093652516868>
https://docs.google.com/document/d/1f_fD92NicBJd93xzFFBiBFzrk8kT9Fdud4rKqF1rbkQ/edit?usp=sharing
meager warren
interesting, if you `with NeoPixel(...) as n: ...`, the pixels all get blanked a...
Adafruit_CircuitPython_NeoPixel/blob/main/neopixel.py has no destructor __del__ method so there won't be anything done when the instance is garbage collected. (I would not recommend adding such a destructor... destruction happens at odd times in an odd order if at all; very non-explicit.)
viscid pine
I couldn't even get __del__ to run for classes defined in python
import gc
class X:
def __del__(self):
print("Del was called")
print("Creating a new X")
my_x = X()
print("Trying to delete X")
del my_x
while True:
gc.collect()
it does seem to run for native classes though
manic glacierBOT
digital shoreBOT
adafruit
teal thorn
Would it make sense to revise the i.MX port to use MCUExpresso?
Thanks for the advice I've made some progress. I decided to start by getting one of the existing boards building from the mcux-sdk. I got it to build, but I still need to test it on hardware.
main furnace
When I run the latest build from S3 (pybadge-en_GB-20211228-ea638c0) os.uname().version thinks it is "7.1.0-rc.0-189". I was expecting it to be "-rc.1".
tulip sleet
When I run the latest build from S3 (pybadge-en_GB-20211228-ea638c0) `os.uname()...
That build is against main, and reflects the 7.1.0-rc.0 tag only because there was a merge from 7.1.x to main, so 7.1.0-rc.0 is the latest tag on main. I am planning to release 7.1.0 final today and then release a 7.2.0-alpha.something, which will make that clearer.
spiral elk
Is the 7.1.0 website update also going to update the tinyuf2 link template to reflect the 0.70 release from yesterday? Because as of now they're still pointing to 0.52, which doesn't exist for some newer boards.
tulip sleet
Is the 7.1.0 website update also going to update the tinyuf2 link template to re...
That is completely separate, and we noticed it last night. We will update it, thanks.
spiral elk
Just wanted to nudge in case it hadn't been noticed
tulip sleet
np, we appreciate the reminder. We were trying to figure out how to automate that, but it's easy to update manually.
manic glacierBOT
manic glacierBOT
tulip sleet
@spiral elk ^^
main furnace
Adafruit CircuitPython 7.1.0 on 2021-12-28; Raspberry Pi Zero 2W with rp3a0
Working over uart. It loads code.py from the SD card! π
tulip sleet
Adafruit CircuitPython 7.1.0 on 2021-12-28; Raspberry Pi Zero 2W with rp3a0
7.1.0 is a branch and does not have the most recent broadcom changes. I hope to release a 7.2.0-alpha.1 very soon.
main furnace
zero2w: Tried USB tether, not working.
@tulip sleet I'll give 7.2.0 a try when it's available. Next up: feathers2.
solar whale
<@!329766224093249548> I'll give 7.2.0 a try when it's available. Next up: feath...
I am not having much luck with current main build for zero2w -- works ok for 4b and cm4io but not for zero2w...at least, not yet....
main furnace
7.1.0 looks good on pybadge.
solar whale
actually -- zero2w boots from the SDCard, but the File system is not accessible....
main furnace
π¦ Wonder how that is different from the 7.1.0 release. I boot from SDCard and have file system access (uart console, not connected via USB otg).
7.1.0 looks good on UnexpectedMaker FeatherS2.
solar whale
π¦ Wonder how that is different from the 7.1.0 release. I boot from SDCard and h...
trying 7.1.0 now
works for me as well -- I think there have been lot of changes in main --- not sure what broke
Also did not like having the USB port connected....
thorny jay
https://blog.adafruit.com/2021/12/28/circuitpython-7-1-0-released/
Oh, I was starting to upgrade all my board to RC1 (I only did FeatherS2) and circup was telling me about a new version C:\Users\MediaMonster>circup list Found device at D:\, running CircuitPython 7.1.0-rc.1. A newer version of CircuitPython (7.1.0) is available. Get it here: https://circuitpython.org/board/unexpectedmaker_feathers2 and I was super confused because I checked twice. So I was in the time frame where the json file was updated but not all the binary available. It is actually a good news as upgrading things like the STM32 is a pain, and I prefer not to do it for RC version.
tulip sleet
the GitHub release happens before the builds are done, which is a couple of hours later. I wish it were smoother, but I'm not sure there's an easy way
thorny jay
I even checked the S3 to see if it was the website not updated.
I got the final hint by checking my email and the message saying you were waiting for the build to finish.
I guess there is no good order and I was just super unlucky but very happy it happend right now.
thorny jay
the GitHub release happens before the builds are done, which is a couple of hour...
Also the Twitter notification use to occurs too early. Maybe it is fixed? I did notice it this time, but maybe the wording could be "Very soon, a new version of CircuitPython will be available" rather than "A new version is available" kind of wording.
Oh the Twitter message was 22 minutes ago. It just say "CircuitPython 7.1.0 Released!"
orchid basinBOT
tulip sleet
Also the Twitter notification use to occurs too early. Maybe it is fixed? I did ...
There are two Twitter notifications: one keys off the initial release, and is automated. The other is generated by publishing blog.adafruit.com blog post. I don't publish the post until circuitpython.org is updated. We could remove the first one.
orchid basinBOT
thorny jay
Upgrade still in progress, maybe circup was confused by my Feather M0 Express board that was still in 6.2.0-rc.0. After upgrading to 7.1.0 everything is normal. Here is the error message: >circup list Found device at D:\, running CircuitPython 6.2.0-rc.0. A newer version of CircuitPython (7.1.0) is available. Get it here: https://circuitpython.org/downloads There was a problem: '6mpy'
tulip sleet
Upgrade still in progress, maybe circup was confused by my Feather M0 Express bo...
upgrade circup: pip3 install --upgrade circup. There aren't new 6.x bundles anymore. The new version of circup will now tell you that it itself needs to be upgraded when necessary.
thorny jay
I should be fine, I totally removed every copy of Python on my PC and re-installed the latest of everything. >circup --version CircUp, A CircuitPython module updater. Version 1.0.3
tulip sleet
@slender iron I am preparing a 7.2.0-alpha.1 release. Merge anything you want to be in it.
slender iron
@tulip sleet I'm just getting my work brain back. A 7.2 alpha is a good idea
onyx hinge
https://pypi.org/project/circuitpython-stubs/#history Iooking at pypi releases of circuitpython-stubs, it appears that they are mis-described. Our CI has versions uploaded from releases and from the latest builds on the main branch. The latest non-released one went in as "7.1.0rc1.dev201" even though it's from main and should say something involving "7.2". Since we use setuptools_scm to get the version number, it's probably the old thing where git describe does surprising things because it uses a heuristic if there are multiple tags reachable from the named commit by different merges: ```shell
$ git fetch --tags origin; git describe --tags origin/main
7.1.0-rc.0-201-g3bc48802b
git describe --first-parent --tags works better right this second, because the 7.1.0-rc.0 tag is visible due to a merge, which makes it not a first-parent of the current tip of master, and there's a sneaky way to modify what setuptools_scm uses. I don't know if this is guaranteed to be the case always, but it seems like it (we create a current tag directly on main, but it gets other commits in its history due to merges only)
this apparently also affects builds of circuitpython, where they show the version string internally: ```shell
$ git checkout origin/main
$ cd ports/unix && make && ./micropython
MicroPython 7.1.0-rc.0-201-g3bc48802b-dirty on 2021-12-28; linux version
Use Ctrl-D to exit, Ctrl-E for paste mode
s3 object names don't give a tag-derived string so they're things like .../adafruit-circuitpython-espressif_esp32s3_devkitc_1-en_US-20211228-3bc4880.bin and it doesn't make a difference (except to humans who want to tell the difference between a 7.1 and a main build from s3)
as a two-step process you can find out what tag is "the highest in version numbering, according to git's internal algorithm" and then get the describe from that: ```
$ BEST_TAG=$(git tag --sort=-v:refname --merged HEAD "[0-9]*" | head -1); git describe --tags --match "$BEST_TAG" HEAD
7.2.0-alpha.0-265-g3bc48802b
That's on me, i hit a problem with uncrustify from apt being to old and I didn't bother fixing it
orchid basinBOT
orchid basinBOT
robust quartz
Hi I made a pull request on the circuitpython-org, please check it. thank you.
onyx hinge
Resolving objects.githubusercontent.com (objects.githubusercontent.com)... 185.199.108.133, 185.199.111.133, 185.199.110.133, ...
Connecting to objects.githubusercontent.com (objects.githubusercontent.com)|185.199.108.133|:443... connected.
HTTP request sent, awaiting response... 503 Egress is over the account limit.``` hum github says github transferred too much data π and a build failed
onyx hinge
we don't have a facility for issuing warnings, do we.....? In this new get_normalized_buffer_raise we could print a warning if the element size wasn't 1 byte, so users have a chance to understand their code will break...
solar whale
Are you warning of an upcoming breaking change?
like a deprecation warning?
I guess I'm asking why this is different from other breaking changes?
tulip sleet
I think deprecations are difficult because they can generate a lot of output, and/or the output will never be seen. I think it's just better to error out with a message. We can do that on a version boundary
onyx hinge
data = array.array('H', range(16))
with ... as spi:
spi.readinto(data, start=3, end=5)
In this case, we're considering changing the meaning of this so that it actually operates on data[3:5] (two 16-bit values) instead of 2 bytes... so there's nothing "to make into an error", unless we want the "data is not a bytes" case to be rejected for a whole version before we enable the new behavior.
onyx hinge
(full disclosure: I'd rather update the docs to say that all these functions use byte indices and ignore the actual data type)
manic glacierBOT
vague thicket
@slender iron may I ask where do you parse (if you do so) config.txt for broadcom port? I would like to look how you do it and try to figure out how to pass display resolution to framebuffer from it, so I can set resolution for HDMI displey. Thanks
timber mango
unfortunately it doesn't work, do you have more examples of how to save files in circuitpython?
manic glacierBOT
timber mango
no unfortunately I added the delay time, and it doesn't record, I think it has to do with the while True, when there is no loop it works, but I need a loop to save the information.
slender iron
<@!252717193496756235> may I ask where do you parse (if you do so) config.txt fo...
CP doesn't parse config.txt
you need to reinit the display to get a different frame size
vague thicket
CP doesn't parse config.txt
Ok, I thought it does as "enable_jtag" was not familiar to me from Raspbian OS. Reinit? How can I do that? Do I need to recompile CP for that or can I do that when CP is running? Thanks π
slender iron
its within CP
I'm looking for an example
first is displayio.release_displays()
import videocore
fb = videocore.Framebuffer(width, height)
import framebufferio
display = framebufferio.FramebufferDisplay(fb)
I think that's right
vague thicket
Ok, many thanks. π I will try that.
onyx hinge
@slender iron looks like broadcom would be good to add to this table https://circuitpython.readthedocs.io/en/latest/README.html#ports
slender iron
<@!252717193496756235> looks like broadcom would be good to add to this table ht...
Yup, I just noticed that too
wraith crow
@slender iron fb=videocore.Framebiffer(60,17) returns ValueError: no fb
wraith crow
LOL oh pixels
slender iron
π
wraith crow
Nice! 1024x480 works on my display
vague thicket
800x480 for me with this (https://www.waveshare.com/wiki/5inch_HDMI_LCD) LCD π IΒ΄m using latest alpha version if 7.2.0 and I noticed that if I have this LCD plug in to the RPi 4B it will not boot up properly and it stuck at rainbow screen. I guess that there is some SPI communication which is blocking boot. Otherwise I have to say, that I did not have problem with SD card corruption. I can eject CP drive, reboot it and it is still working.
manic glacierBOT
ornate breach
hey @tulip sleet are you around?
tulip sleet
yes, though not for long π€
ornate breach
have you had any issues building ESP32-S2 boards?
I've been getting issues on the auto_display_resources part of the build
i'll grab you the exact error i'm getting
tulip sleet
nope, as long as espressif/esp-idf is up to date. We changed versions around the 7.1.x/main split, so did you update the submodules?
ornate breach
i did
tulip sleet
ga and paste the error
ornate breach
i merged in latest on my circuitpython fork and pulled that down
../../extmod/ulab/code/numpy/approx/approx.c: In function 'approx_interp':
../../extmod/ulab/code/numpy/approx/approx.c:61:24: error: too few arguments to function 'ndarray_from_mp_obj'
ndarray_obj_t *x = ndarray_from_mp_obj(args[0].u_obj);
^~~~~~~~~~~~~~~~~~~
In file included from ../../extmod/ulab/code/numpy/approx/../../ulab_tools.h:14,
from ../../extmod/ulab/code/numpy/approx/approx.c:21:
../../extmod/ulab/code/numpy/approx/../../ndarray.h:201:16: note: declared here
ndarray_obj_t *ndarray_from_mp_obj(mp_obj_t , uint8_t );
^~~~~~~~~~~~~~~~~~~
../../extmod/ulab/code/numpy/approx/approx.c:62:25: error: too few arguments to function 'ndarray_from_mp_obj'
ndarray_obj_t *xp = ndarray_from_mp_obj(args[1].u_obj); // xp must hold an increasing sequence of independent values
^~~~~~~~~~~~~~~~~~~
from ../../extmod/ulab/code/numpy/approx/approx.c:21:
../../extmod/ulab/code/numpy/approx/../../ndarray.h:201:16: note: declared here
ndarray_obj_t *ndarray_from_mp_obj(mp_obj_t , uint8_t );
^~~~~~~~~~~~~~~~~~~
../../extmod/ulab/code/numpy/approx/approx.c:63:25: error: too few arguments to function 'ndarray_from_mp_obj'
ndarray_obj_t *fp = ndarray_from_mp_obj(args[2].u_obj);
^~~~~~~~~~~~~~~~~~~
In file included from ../../extmod/ulab/code/numpy/approx/../../ulab_tools.h:14,
from ../../extmod/ulab/code/numpy/approx/approx.c:21:
../../extmod/ulab/code/numpy/approx/../../ndarray.h:201:16: note: declared here
ndarray_obj_t *ndarray_from_mp_obj(mp_obj_t , uint8_t );
^~~~~~~~~~~~~~~~~~~
../../extmod/ulab/code/numpy/approx/approx.c: In function 'approx_trapz':
../../extmod/ulab/code/numpy/approx/approx.c:160:24: error: too few arguments to function 'ndarray_from_mp_obj'
ndarray_obj_t *y = ndarray_from_mp_obj(args[0].u_obj);
from ../../extmod/ulab/code/numpy/approx/approx.c:21:
../../extmod/ulab/code/numpy/approx/../../ndarray.h:201:16: note: declared here
ndarray_obj_t *ndarray_from_mp_obj(mp_obj_t , uint8_t );
^~~~~~~~~~~~~~~~~~~
../../extmod/ulab/code/numpy/approx/approx.c:177:13: error: too few arguments to function 'ndarray_from_mp_obj'
x = ndarray_from_mp_obj(args[1].u_obj); // x must hold an increasing sequence of independent values
^~~~~~~~~~~~~~~~~~~
In file included from ../../extmod/ulab/code/numpy/approx/../../ulab_tools.h:14,
from ../../extmod/ulab/code/numpy/approx/approx.c:21:
../../extmod/ulab/code/numpy/approx/../../ndarray.h:201:16: note: declared here
ndarray_obj_t *ndarray_from_mp_obj(mp_obj_t , uint8_t );
^~~~~~~~~~~~~~~~~~~
make: *** [build-odt_branch-esp32s2/extmod/ulab/code/numpy/approx/approx.o] Error 1```so ulab submodule is not up to date?
tulip sleet
are you at the tip of main?
tulip sleet
ok, trying it myself
ornate breach
i'm on mac. i've had issues with it taking a lot of syncing submodules and init'ing them to get everything to work the last couple weeks. might be my computer
tulip sleet
cd circuitpython/extmod/ulab, and do git status and let's look at the commit
halbert@tuna:~/repos/adafruit/circuitpython/extmod/ulab$ git status
HEAD detached at a99e0b9
nothing to commit, working tree clean
Untracked files:
(use "git add <file>..." to include in what will be committed)
code/fft/
code/numpy/approx/
code/numpy/compare/
code/numpy/filter/
code/numpy/numerical/
code/numpy/poly/
code/numpy/stats/
code/numpy/transform/
code/numpy/vector/
docs/manual/autoapi/
docs/manual/source/ulab/
tests/common/
tests/numpy/
tests/scipy/
tests/utils/
nothing added to commit but untracked files present (use "git add" to track)```same commit but it's showing untracked files
tulip sleet
this is on a clone of adafruit/circuitpython, not my own fork
what is git --version ?
ornate breach
2.23
tulip sleet
I am at 2.34.1, and there is junk in your ulab.
I would suggest cd circuitpython; git clean -xxffdd to remove all the non-repo stuff (stash any uncommited work), then see if that makes any difference. Or reclone your fork from scratch
a bit odd that the commit is the same but yours is showing an extra digit, but maybe that is just a cosmetic git difference
my build worked fine. You could try a fresh adafruit/circuitpython clone and see if it works
ornate breach
do i need to reinit the submodules after git clean -xxffdd
tulip sleet
I don't think so
ornate breach
i'll try these things and hopefully find success
manic glacierBOT
manic glacierBOT
timber mango
Thanks for helping me dahn π
manic glacierBOT
manic glacierBOT
manic glacierBOT
thorny jay
Ideas and thinks that could be fixed about mpy:
- https://circuitpython.readthedocs.io/en/latest/docs/troubleshooting.html#valueerror-incompatible-mpy-file say:
In particular, the mpy binary format changed between CircuitPython versions 1.x and 2.x, 2.x and 3.x, and will change again between 6.x and 7.x.
We should change the "and will change again" as this is done already. - While the CircuitPython_Bundle has been released recently and only contain 7.x mpy (and no more 6.x).
The community bundle has not been updated since November 27 and contain both 7.x and 6.x mpy. Maybe it need a new release or there is a configuration difference to fix.
manic glacierBOT
manic glacierBOT
2340.0 cannot be represented exactly with the bits we have. See https://www.h-schmidt.net/FloatConverter/IEEE754.html
Note that the float representation we use drops the two lowest bits as well: it's a 30-bit float with two bits less of mantissa. There is more discussion of this kind of thing in the MicroPython issues, for instance: https://github.com/micropython/micropython/issu...
crimson ferry
does CircuitPython only present MSC at start-up? Is there a way to mount CIRCUITPY (either from the Mac, or from code without hard reset) if it's been ejected (macOS)?
CDC data (console / REPL) is available, but device doesn't show up in Disk Utility
tulip sleet
does CircuitPython only present MSC at start-up? Is there a way to mount CIRCUIT...
I think the USB devices may need to re-enumerate. It might be possible to do some kind of suspend/resume via some calls to TinyUSB, but I am not sure.
crimson ferry
thanks, Iβll take a look
tough flax
Thank you, #circuitpython-dev for everything you've done in 2021 to help us who make custom AT solutions
solar whale
Thank you, <#327298996332658690> for everything you've done in 2021 to help us w...
Thank you for making all this important.
manic glacierBOT
onyx hinge
oooh I have now officially written an asyncio program. It shows a blinking "12:00" potentially on multiple of the ht16k33 displays each one at a slightly different rate. Use address jumper A2 to tell CircuitPython if a display is 7-segment. https://gist.github.com/6fe570c0bcd9560a9a7822bdd58fc648
manic glacierBOT
dusky halo
Question - for pioasm (or general circuitpy dev), I see both RuntimeError and SyntaxError for PIO asm. Is there a guideline anywhere on which to use and when? So my use case is catching invalid cases (which I found myself :D) of set registers, or mov operations missing spaces, and coming out with nicer exceptions (see https://github.com/adafruit/Adafruit_CircuitPython_PIOASM/pull/32).
digital shoreBOT
adafruit
onyx hinge
@dusky halo I don't see any particular logic to the existing SyntaxError vs RuntimeError exceptions in pioasm.
@dusky halo raise RuntimeError("Invalid mov source:", source) from exc any particular reason this wasn't an f""-string? Is this two argument RuntimeError correct?
tulip sleet
I think ValueError might be more appropriate for an invalid operand. I would reserve SyntaxError for what it says, invalid syntax (unparseable input line)
RuntimeError is usually used for unexpected things beyond the user's control, not programming errors.
we use ValueError all the time time for say, lengths < 0, etc.
dusky halo
<@!693802301575528509> `raise RuntimeError("Invalid mov source:", source) from e...
No reason really - probably being inconsistent with it. It logs intelligibly, but I'm not wedded to either syntax.
@tulip sleet that sounds reasonable. For now I'll modify those that I've added - but do we have a good place in the repo to record a dev guideline like that? Like a dev.md or something?
Ok tidied up my PR a bit - made the exceptions consistent, and added tests for my changes.
jaunty juniper
ancient whale
is the version of uncrustify in the debian repos suitable for building circuitpython?
(Uncrustify-0.72.0_f)
viscid pine
on ubuntu i had success using the ppa: https://learn.adafruit.com/building-circuitpython/linux
ancient whale
right, i saw that
but i'm on debian
and i was wondering if the version in the normal repos will work
looks like it's more up to date
it should hopefully work
viscid pine
i tried using a more up to date one at first and it didn't work out
0.71 is what the ppa gives me, you might be ok
actually if you're just building you don't need it
viscid pine
it's for formatting
ancient whale
i know
onyx hinge
@ancient whale I usually use the version in debian bullseye (0.72.0+dfsg1-2) and usually get by without problem, fwiw.
manic glacierBOT
ancient whale
i'm getting an error```
~/circuitpython/ports/atmel-samd$ make BOARD=adafruit_neokey_trinkey_m0 V=2
GEN build-adafruit_neokey_trinkey_m0/genhdr/mpversion.h
python3 ../../py/makeversionhdr.py build-adafruit_neokey_trinkey_m0/genhdr/mpversion.h
Creating build-adafruit_neokey_trinkey_m0/frozen_mpy.c
python3 ../../tools/mpy-tool.py -mlongint-impl=none -f -q build-adafruit_neokey_trinkey_m0/genhdr/qstrdefs.preprocessed.h > build-adafruit_neokey_trinkey_m0/frozen_mpy.c
usage: mpy-tool.py [-h] [-d] [-f] [--merge] [-q QSTR_HEADER]
[-mlongint-impl {none,longlong,mpz}] [-mmpz-dig-size N]
[-o OUTPUT]
files [files ...]
mpy-tool.py: error: the following arguments are required: files
make: *** [../../py/mkrules.mk:167: build-adafruit_neokey_trinkey_m0/frozen_mpy.c] Error 2
make: *** Deleting file 'build-adafruit_neokey_trinkey_m0/frozen_mpy.c'
stuck elbow
@ancient whale did you upgrade git submodules?
ancient whale
<@!642358695652753418> did you upgrade git submodules?
i used make fetch-submodules
stuck elbow
do you have any files in, say, frozen/Adafruit_CircuitPython_NeoPixel?
ancient whale
yep
~/circuitpython$ ls frozen/Adafruit_CircuitPython_NeoPixel
CODE_OF_CONDUCT.md examples LICENSES README.rst requirements.txt
docs LICENSE neopixel.py README.rst.license setup.py
If you did a build before you had your submodules, you may be able to fix the problem by make BOARD=... clean
where ... is the board you're building for
ancient whale
i think it's working
solar whale
I always do make ... clean before any build.
ancient whale
Yay! It worked! Thanks
alright now that i have the build working, how can i add a builtin native module
ok c modules are in extmod
ancient whale
oo there's micropython library docs https://docs.micropython.org/en/latest/develop/library.html
tulip sleet
@mortal mica reminder to PR your new pykey boards to circuitpython-org
upbeat plover
why is my username highlighted? did i do something wrong?
idle owl
why is my username highlighted? did i do something wrong?
It thinks it's tagging you, so it highlights your name. Is that on GitHub?
You did nothing wrong.
upbeat plover
yeah under releases, thank you for very fast reply.
idle owl
No problem! It sees it as a tag event, so it highlights your name.
@onyx hinge Are you around for a quick discussion about your running the meeting suggestions?
orchid basinBOT
onyx hinge
@idle owl text is better right now but otherwise yes
idle owl
<@330227457296957440> text is better right now but otherwise yes
Text is fine. I'm happy with most of it, but have a question about one thing. You suggested deleting the entire bit about the meeting being recorded because it's mentioned pre-meeting, I don't see it there.
I agree it's obvious we recorded the meeting and isn't worth mentioning.
But it should be mentioned before.
onyx hinge
My goals were to make the calendar more prominent and to change some in meeting stuff too be more directed at later viewers then live participants
idle owl
You did a great job. That's what I thought you were aiming for.
I also think we should keep something about the "if you can't participate during, but want to be a part of it, leave notes for us to read"
onyx hinge
Since we almost always gloss over the pre meeting section I probably don't know what's there and I didn't check
idle owl
because later listeners might be an audience for that concept, since they didn't attend
idle owl
Since we almost always gloss over the pre meeting section I probably don't know ...
Fair enough.
I'll paste that text into the pre-meeting stuff.
onyx hinge
Makes sense, thanks for looking it over
idle owl
@onyx hinge Do you want to reword the bit about adding your notes if you can't attend to make it shorter?
onyx hinge
Do you want me to do a back up recording today?
idle owl
I approved everything else.
idle owl
Do you want me to do a back up recording today?
OMG yes. Please please please.
I am still recovering from booster, for one, and two, this will be the first actual recording with this new laptop.
So.... I expect problems.
idle owl
Makes sense, thanks for looking it over
I updated the text and added it to the end of the "notes doc" section. Anyway. Thank you for going through that and tightening it up.
onyx hinge
@idle owl whoops got distracted
idle owl
<@!330227457296957440> whoops got distracted
No worries. π
onyx hinge
I will do backup recording!
idle owl
Thank you so much
onyx hinge
I lost a component right off my desk and had to call in Ingrid to find it for me because losing things within arm's reach is my super power
idle owl
<@&356864093652516868> Hello, and happy new year! We have our first CircuitPython Weekly meeting of 2022 in about 1 hour 45 minutes. If you plan to participate, whether or not you're attending, please add your Hug Reports and Status Updates to the notes doc. If you are simply planning to listen in, no action is needed. Looking forward to chatting with all of you soon! https://docs.google.com/document/d/1f_fD92NicBJd93xzFFBiBFzrk8kT9Fdud4rKqF1rbkQ/edit?usp=sharing
thorny jay
<@&356864093652516868> Hello, and happy new year! We have our first CircuitPytho...
Hi, in my notes I have put optional extra written information that does not need to be read-out. Those are extra details or notes. I used parenthesis to separate that from the main point. The idea is to have it saved if someone want to dig more, but not to take too much reading time for you. Does that make sense or should I try to rewrite or just remove that? Or any preferred special tagging like ((( extra bla bla bla ))).
idle owl
Hi, in my notes I have put optional extra written information that does not need...
It made sense! I don't mind reading all of it, if you'd like. But I definitely understood what you were going for there.
thorny jay
I am usually after Dan, so you can have a pre-read and decide if it is worst reading or not.
idle owl
Will do!
thorny jay
Thanks.
Someone has pasted ```raspberrypi_cm4
espressif_esp32s3_box
espressif_esp32s3_devkitc_1
espressif_esp32s3_devkitc_1_nopsram
bluemicro833
Seeed_XIAO_nRF52840_Sense
jpconstantineau_pykey18
jpconstantineau_pykey87
pimoroni_tiny2040_2mb
jpconstantineau_pykey44
seeeduino_xiao_rp2040
odt_cast_away_rp2040
sparkfun_stm32_thing_plus#
Could be new supported board in 7.2.0-alpha.1
idle owl
Hmm. I'll take a look.
I deleted it. I didn't include the extensive details because there's so many updates.
blissful pollen
Alright, who wants to do an audio check with me?
if you need someone im here
idle owl
Perfect, thank you!
turbid radish
That list is the boards added without corresponding notes at circuitpython.org so there are a bunch of "Unknown Board" entries
Sunny and 66 F here
blissful pollen
stopped being -20 F here, now around 0 F
turbid radish
π
lone axle
solar whale
yes!
lone axle
There are a few display drivers that have non-Displayio and Displayio ones seperately. The ones I've seen tend to be much older frambuffer based libraries. Most of the newer hardware has just Displayio I think though.
errant grail
Thanks!
idle owl
No crashes!
onyx hinge
yay @idle owl
idle owl
Now to find out if the audio worked.
idle owl
I'm headed out for a bit, π
I'll let you know if I need your recording.
Dropping off to check audio. Thanks everyone!
idle owl
ah I'll wait until you have checked that
Sounds ok? without listening to the entire thing.
mortal mica
<@!347768391773454337> reminder to PR your new pykey boards to circuitpython-org
Thanks for the nudge. I did a PR a while ago (https://github.com/adafruit/circuitpython-org/pull/815) but I left 2 of them "hidden" as they are not live yet on Tindie. Now that I am back from holidays, I'll need to get it in before I go back to work tomorrow (my day job...)
idle owl
Here is the notes document for Mondayβs 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 youβll be attending the meeting - itβs super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and weβll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/1f-olt8jHqYvCX-7hbH1kZF7qILHwaqk1mgiOKxqVhV8/edit?usp=sharing
mortal mica
@idle owl The video of the meeting just showed up on youtube but the link to the notes seems to be missing. Happy New Year eneryone!
idle owl
<@!330227457296957440> The video of the meeting just showed up on youtube but th...
Oof. Thanks for catching that!
idle owl
<@!330227457296957440> The video of the meeting just showed up on youtube but th...
Refresh?
mortal mica
wow our messages must have crossed paths on their way...
idle owl
No, the link I posted above is for next week.
I forgot to add the link for this week to this week's video.
mortal mica
It's up now.
thorny jay
Did someone recently try using RPi imager to write an image for the Pi2W? I'll try older version, but it could be related to the fact that the image first action is to format the CPYTHON drive. I remember Danh saying something about a new behaviour for the general CP when that drive is missing. So maybe it interact badly? I reverted to 7.1.0 and it works... but maybe I waited longer. But there might be something fishy, that you don't notice if you only update kernel8.img?
thorny jay
I start to be confident that adafruit-circuitpython-raspberrypi_zero2w-en_US-20220102-f2f38c7.disk.img and adafruit-circuitpython-raspberrypi_zero2w-en_US-7.2.0-alpha.1.disk.img are broken while adafruit-circuitpython-raspberrypi_zero2w-en_US-7.1.0.disk.img is working OK. It might be a timing issue, but I waited enough for the drive to be created at first startup. Sometime I see a console, and those error message including the above. I tried just pushing adafruit-circuitpython-raspberrypi_zero2w-en_US-7.2.0-alpha.1.kernel8.img and adafruit-circuitpython-raspberrypi_zero2w-en_US-20220102-f2f38c7.kernel8.img to a previously 7.1.0 image and it fail on me too. I have all kind of USB error message, my Win10 PC does not like what I do.
proven garnet
Hey @thorny jay, just wanted to check in about this issue you had that might be taken care of, aiming for the sweet, sweet Issue Closed if applicable 
https://github.com/adafruit/Adafruit_CircuitPython_BLE_BroadcastNet/issues/13
thorny jay
Didnt' you add a few line to support that, like one or two weeks ago?
Those did seems to make a lot of sense, but I did not tested.
proven garnet
Yeah, the issue is still open though so I didn't know if it had been addressed
thorny jay
I believe so.
proven garnet
Or whether I just add some fun new characters to hang out in the source code haha
thorny jay
Not sure many use that ble_broadcast nowdays.
It is rather complex, you need a Pi and a CircuitPython board, and Adafruit.io ... that is a lot I guess.
It could be concept done before the ESP32S2 support.
proven garnet
It's the CPB right? Was looking to get one, and I managed to snag the RPi 2 Zero W before the holidays, and for the trifecta I've been meaning to checkout Adafruit.io. Maybe I'll do that once I get the Pi up
Actually, on that note, I can let you know when I do if you want me to try out anything regarding what you wrote above.
thorny jay
Don't hesitate.
A ping here is more efficient that the notification by email that Github will send me... I don't check emails every day anymore.
trim elm
Hey @lone axle I was doing some sweeps, should this be in the bundle? https://github.com/adafruit/Adafruit_CircuitPython_24LC32
lone axle
Hey <@!382939733107408897> I was doing some sweeps, should this be in the bundle...
Yes, but ideally this PR for it will be merged https://github.com/adafruit/Adafruit_CircuitPython_24LC32/pull/6 and the initial release can be made for it first. After that it should be good to go for adding to the bundle.
That PR fixes many of the docstrings and the example get everything working with the latest version of the library. So ideally we'll get that in before adding to the bundle to avoid anyone getting confused by the current state.
trim elm
ah ok. I'll take a look at it and see if I can approve it myself and forward it to someone who can if I don't know enough to
lone axle
Thank you!
spiral elk
Yes, but ideally this PR for it will be merged https://github.com/adafruit/Adafr...
Given that it will drive pretty much any double-byte addressed EE (and expect a PR to support single-byte addresses soon) is the name really the most accurate?
lone axle
Given that it will drive pretty much any double-byte addressed EE (and expect a ...
We discussed it a bit during the meeting (not today, but recently). The name was chosen so that it matches specific breakout devices to make it easy for folks to know which library to use.
spiral elk
I was planning on working on the single-byte PR over the holidays but didn't get around to it.
lone axle
In this case the driver can support other devices with relatively minor changes (mostly updating the max size value). But this specific repo will remain for the 24LC32 Adafruit breakout.
manic glacierBOT
manic glacierBOT
orchid basinBOT
pallid echo
During Christmas break I got the main branch of CircuitPython working on this cute little Waveshare RP2040-Zero board. The onboard neopixel is happily cycling through hsl hues using the neopixel library and board.NEOPIXEL. So I know I got the most important things in the board definition correct 
. For the USB_VID and USB_PID I jotted down the numbers from the firmware it came with that showed up with lsusb. I doubt those are correct; the string said "RPI-RP2" and the VID is from the Raspberry Pi Foundation... But then I saw other board definitions in the tree with the same VID. And now I'm a bit confused.
Hmmm. That looks like a statement, but feels like a question. π I guess my question is: what do I need to research or ask the manufacturer in order to get this closer to a PR?
idle owl
Hmmm. That looks like a statement, but feels like a question. π
I guess my ques...
I feel like there's a thing with the RP2040, that it has a set VID or some such. But others will know better than me. Folks will be around soon.
pallid echo
I feel like there's a thing with the RP2040, that it has a set VID or some such....
π Thanks. It was surprisingly smooth to get built. I don't know who to compliment for that. But I only just now found https://learn.adafruit.com/how-to-add-a-new-board-to-circuitpython/customizing-the-board-files and I got this working first try. So the project layout and code organisation is pretty good.
idle owl
π Thanks. It was surprisingly smooth to get built. I don't know who to complime...
Great to hear! The whole team deserves the compliment for that.
orchid basinBOT
idle owl
@gilded cradle Does that Blinka_DisplayIO PR need testing?
slender iron
@pallid echo the PID will be a number. "RPI-RP2" is the drive name. You can get a PID for the board you added here: https://github.com/raspberrypi/usb-pid
gilded cradle
@idle owl I did some basic testing with the BrainCraft and the basic demo script, but nothing extensive, so some additional testing would probably be helpful.
pallid echo
<@!372766782030741504> the PID will be a number. "RPI-RP2" is the drive name. Yo...
Thanks! I'll point Waveshare to the form.
gilded cradle
I'll make a note of that in the PR
idle owl
I'll make a note of that in the PR
Thanks!
slender iron
@onyx hinge have you seen weblate hit a rate limit?
any folks with recent SD card experience? https://forums.adafruit.com/viewtopic.php?f=60&t=186753
onyx hinge
@slender iron no -- I don't know why that would happen
manic glacierBOT
onyx hinge
@slender iron That seems to be a message github is sending back to weblate, it might happen if they had a bug or configuration error that made them make too many github API requests
I don't see any opened issue in their github about it
manic glacierBOT
slender iron
@onyx hinge looks like it's working again. I told it to push through the weblate UI
I just needed to go on a run first π
timber mango
Hi,
I need to set CFG_TUSB_MEM_ALIGN defined in supervisor/shared/usb/tusb_config.h to custom value (256) in order to get USB running on NXP LPC55Sxx chips.
Tried to set the CFG_TUSB_MEM_ALIGN in nxp/mpconfigport.h
Modifing the tusb_config.h in following way.
#if !defined(CFG_TUSB_MEM_ALIGN) #define CFG_TUSB_MEM_ALIGN __attribute__ ((aligned(4))) #endif
But the problem is that tusb_config.h is included directly in lib/tinyusb/src/device/usbd.h
Any suggestion how to overcome this issue?
Is it safe to simply change the value from 4 to 256?
tulip sleet
Hi,
I need to set CFG_TUSB_MEM_ALIGN defined in supervisor/shared/usb/tusb_conf...
There is a supervisor/shared/usb/tusb_config.h. You can add coditional preprocessor code to that. Note, for instance lines like:
#define CFG_TUD_MSC CIRCUITPY_USB_MSC which are already in there
oh never mind, you already said that
but what i'm saying is, you can put some dependencies from outside in there
like make the 4 be the default, and create a new variable that lets you override the 4
#define CFG_TUSB_MEM_ALIGN __attribute__ ((aligned(4)))
to
#ifndef CIRCUITPY_TUSB_MEM_ALIGN
#define CIRCUITPY_TUSB_MEM_ALIGN (4)
#endif
#define CFG_TUSB_MEM_ALIGN __attribute__ ((aligned(CIRCUITPY_TUSB_MEM_ALIGN)))
would that work, or is it the same problem?
timber mango
would that work, or is it the same problem?
I am afraid that the problem is that tusb_config.h is included by tinyusb directly.
There is no chance to set CIRCUITPY_TUSB_MEM_ALIGN from circuitpython this way.
Here is the place in lib/tinyusb/src/tusb_option.h where this file is included.
// Allow to use command line to change the config name/location
#ifdef CFG_TUSB_CONFIG_FILE
#include CFG_TUSB_CONFIG_FILE
#else
#include "tusb_config.h"
#endif
I'll try is to override the file name (CFG_TUSB_CONFIG_FILE).
tulip sleet
I am afraid that the problem is that tusb_config.h is included by tinyusb direct...
But the include path for tusb_config.h is controlled by the build, and we do send things into that file, like CIRCUITPY_USB_MSC. CIRCUITPY_USB_MSC is actually defined in various *.mk files, and then is converted to a C definition in py/circuitpy_mpconfig.mk, like this:
CIRCUITPY_USB_MSC ?= $(CIRCUITPY_USB)
CFLAGS += -DCIRCUITPY_USB_MSC=$(CIRCUITPY_USB_MSC)
so you should be able to do this
add CIRCUITPY_TUSB_MEM_ALIGN = 256 to the appropriate mpconfigport.mk or mpconfigboard.mk, and then add the -DCIRCUITPY_TUSB_MEM_ALIGN stuff to circuitpy_mpconfig.mk, analogous to the above.
ifdef CIRCUITPY_USB_MSC
CFLAGS += -DCIRCUITPY_TUSB_MEM_ALIGN=$(DCIRCUITPY_TUSB_MEM_ALIGN)
Then in tusb_config.h, as above:
#ifndef CIRCUITPY_TUSB_MEM_ALIGN
#define CIRCUITPY_TUSB_MEM_ALIGN 4
#endif
#define CFG_TUSB_MEM_ALIGN __attribute__ ((aligned(CIRCUITPY_TUSB_MEM_ALIGN)))
Sounds good. Let me try it out.
timber mango
Then in `tusb_config.h`, as above:
```c
#ifndef CIRCUITPY_TUSB_MEM_ALIGN
#define...
Indeed it worked out. Thank you for the hint!
@tulip sleet The pull request (https://github.com/adafruit/circuitpython/pull/5453) is pending since quite a long time. Actually it became obsolete, because reworked the code significantly in the meantime.
I am going to cancel the PR#5453 and create new one with the latest changes.
Is it fair enough?
tulip sleet
<@!329766224093249548> The pull request (https://github.com/adafruit/circuitpyt...
That's perfectly fine. A clean PR is nice.
timber mango
Another question regarding the ringbuffer.
I would like to add new function there to allow user buffer allocation:
bool ringbuf_init(ringbuf_t *r, uint8_t *buf, size_t capacity) { r->dynamic = false; r->buf = buf; r->size = capacity + 1; r->iget = r->iput = 0; return r->buf != NULL; }
So far the static allocation looked like this:
// Static initialization: // byte buf_array[N]; // ringbuf_t buf = {buf_array, sizeof(buf_array)};
This approach exposes the order of the fields in ringbuffer structure.
By using dedicated function (ringbuf_init), which still backward compatible with static and dynamic allocation, this risk is avoided.
However I had to add new field to ringbuffer structure:
typedef struct _ringbuf_t { uint8_t *buf; // Allocated size; capacity is one less. Don't reference this directly. uint32_t size; uint32_t iget; uint32_t iput; bool dynamic; } ringbuf_t;
Does it make sense?
tulip sleet
do you test dynamic somewhere?
timber mango
do you test `dynamic` somewhere?
I believe, yes. But I need to double check it.
tulip sleet
ringbuf is a low-overhead thing from MicroPython. It could be made more general in a bunch of ways
I am not all that happy with it - for instance the 2-byte stuff is somewhat hacky
but it is fast and small
you would only test dynamic on freeing or something, is that right?
is it heap-allocated? You will need to worry about garbage collection and making the pointer be reachable.
timber mango
ringbuf is a low-overhead thing from MicroPython. It could be made more general ...
MPY uses a bit different function set. So the ringbuffer structure is CPY specific if I am not wrong.
tulip sleet
Yes, I cleaned it up a bit a while ago
timber mango
you would only test `dynamic` on freeing or something, is that right?
Right.
`// Dynamic initialization. This should be accessible from a root pointer.
// capacity is the number of bytes the ring buffer can hold. The actual
// size of the buffer is one greater than that, due to how the buffer
// handles empty and full statuses.
bool ringbuf_alloc(ringbuf_t *r, size_t capacity, bool long_lived) {
r->dynamic = true;
r->buf = gc_alloc(capacity + 1, false, long_lived);
r->size = capacity + 1;
r->iget = r->iput = 0;
return r->buf != NULL;
}
void ringbuf_free(ringbuf_t *r) {
if (r->dynamic) gc_free(r->buf);
r->buf = (uint8_t *)NULL;
r->size = 0;
ringbuf_clear(r);
}`
timber mango
ringbuf is a low-overhead thing from MicroPython. It could be made more general ...
You know, perhaps it's a good idea to talk to MPY folks in this respect as well.
tulip sleet
I think gc_free will notice that the pointer is outside of the heap and not do anything, so you don't need to guard gc_free() that way
timber mango
I think gc_free will notice that the pointer is outside of the heap and not do a...
dynamic is set to false, when static allocation (ringbuf_init called) is used and set to true in case of dynamic allocation (ringbuf_alloc is called). Thus gc_free() is will not be called with statically allocated buffer at all.
tulip sleet
what I said above is wrong, I think it checks that the pointer is in the heap and asserts out otherwise, but in general, for a particular ring buffer, you would know whether it is allocated dynamically or not, and wouldn't call ringbuf_free() on a non-heap ringbuf. What is your use case here?
you could do:
ringbuf_t buf = {
.buf = ...,
.size = ...,
};
and then not worry about the order
timber mango
you could do:
```c
ringbuf_t buf = {
.buf = ...,
.size = ...,
};
```
It's true. But in this case user code needs to know the names of the structure fields.
timber mango
what I said above is wrong, I think it checks that the pointer is in the heap an...
The use case is the following: DEBUG_TX/DEBUG_RX are set (e.g. Debug UART interface is used) and the UART from common-hal.busio uses the ringbuffer structure. Then Debug UART instance is initialised with static buffer before VM heap is initialised.
tulip sleet
The use case is the following: DEBUG_TX/DEBUG_RX are set (e.g. Debug UART interf...
so that would be OK, or you could use an external boolean that indicates that particular ring buffer is on the heap. But a single boolean is small. Maybe rename it to heap from dynamic, because the buffer itself cannot be changed dynamically.
manic glacierBOT
manic glacierBOT
crimson ferry
ESP32-S2 TFT Feather just hit the shop https://www.adafruit.com/product/5300
solar whale
Why do new products always show up within hours of my placing an order? π
slender iron
they are also hours before AAE π
crimson ferry
Why do new products always show up within hours of my placing an order? π
yup, I also get too nervous that it will get sold out before the code during the show
blissful pollen
@idle owl I tried the IS31 code on the QtPy M0 Haxpress and seemed to still work with the additions. Only some of the samples would run with or without due to memory limitations. I doubt it will run on anything smaller then that.
idle owl
<@!330227457296957440> I tried the IS31 code on the QtPy M0 Haxpress and seemed ...
Ok. Did you attempt to use it in the current state? i.e. without your updates?
blissful pollen
Ok. Did you attempt to use it in the current state? i.e. without your updates?
Yeah I tried it in both. Most of the guide examples won't work either way but was able to run the fire eyes example
idle owl
Ok, that's actually the result I was hoping for.
So let's include your code in the current library. Obviously make sure it remains backwards compatible.
blissful pollen
You have to use the mpy compiled version too it is that tight on memory. Using the python version ran out
idle owl
But let's make it work with what already has it.
Python version always runs out π
That's not a surprise.
blissful pollen
I'll try to take a look at it this weekend if I don't get some time before
idle owl
Perfect. Thanks for your patience with this decision. I really appreciate it!
blissful pollen
No problem I was in no rush myself, thanks for your help in looking at it too
tulip sleet
<@!329766224093249548> <@!426200547024961539> do you know how to disable `__exid...
i don't even know what it is, sorry
slender iron
something for exception unwinding
onyx hinge
https://stackoverflow.com/questions/9752000/exidx-start-and-exidx-end-what-do-they-do so they have to do with C++ exceptions -- is there C++ source in there?
-fno-exceptions?
slender iron
I think I just tried -fno-exceptions
it's just the dfu and usbtmc examples that fail. I think I'll just skip them
onyx hinge
/home/runner/cache/toolchain/xpack-arm-none-eabi-gcc-10.2.1-1.1/bin/../lib/gcc/arm-none-eabi/10.2.1/../../../../arm-none-eabi/bin/ld: /home/runner/cache/toolchain/xpack-arm-none-eabi-gcc-10.2.1-1.1/bin/../lib/gcc/arm-none-eabi/10.2.1/libgcc.a(unwind-arm.o): in function `get_eit_entry': so the start files are referring to this function...
you'd get to trace back why unwind-arm is linked in in the first place, not fun
slender iron
it builds for me with gcc 11
onyx hinge
it's both host examples yeah?
slender iron
no, device
onyx hinge
oh there's a table at the bottom, must be incomplete or something
I saw 4 rows and 2 failures and drew a wrong conclusion
Build Summary: 16 succeeded, 0 failed, 7 skipped and took 4.06s .. with gcc version 10.2.1 20201103 (release) (xPack GNU Arm Embedded GCC, 64-bit) locally
so that's weird
oh I got it after you added the skips
slender iron
ah π
onyx hinge
how can I get build_family.py to show me what it actually invokes? or to invoke make V=2 or whatever the magic is
slender iron
Β―_(γ)_/Β―
onyx hinge
/home/jepler/src/tinyusb/xpack-arm-none-eabi-gcc-10.2.1-1.1/bin/../lib/gcc/arm-none-eabi/10.2.1/../../../../arm-none-eabi/bin/ld: /home/jepler/src/tinyusb/xpack-arm-none-eabi-gcc-10.2.1-1.1/bin/../lib/gcc/arm-none-eabi/10.2.1/libgcc.a(unwind-arm.o): definition of __aeabi_unwind_cpp_pr0
```for some reason the standard C library functions like strtol refer to this `__aeabi_unwind_cpp_pr0` symbol, which brings in `unwind-arm.o`huh, interesting
onyx hinge
I had a longer commit message which git ate and wouldn't give back. It explained that these symbols create an empty exception information table. The table is required because of a dependency we don't understand from C routines like strtol, which refers to the symbol __aeabi_unwind_cpp_pr0. That brings in get_eit_entry from the same object file, which needs __exidx_start and __exidx_end. It would be best if these routines weren't brought in, but we can't control that, and providing an empty table seems like the next best thing.
The link.ld file is copied from the bsp with just the two symbols added.
robust quartz
Hi I compile the Seeed_XIAO_nRF52840_Sense board is work on my local, coz I add the SOFTDEV_VERSION=7.0.1 in behind the "make BOARD=...SOFTDEV_VERSION=7.0.1", I think I downloaded that board from the Circuitpython web one does not add "SOFTDEV_VERSION=7.0.1", It cause the start address wrong.
I think the system build the uf2 is default make BOARD=..., how can I make it to add SOFTDEV_VERSION=7.0.1 behind the make BOARD=...
manic glacierBOT
Hi I compile the Seeed_XIAO_nRF52840_Sense board is work on my local, coz I add the SOFTDEV_VERSION=7.0.1 in behind the "make BOARD=...SOFTDEV_VERSION=7.0.1", I think I downloaded that board from the Circuitpython web one does not add "SOFTDEV_VERSION=7.0.1", It cause the start address wrong.
I think the system build the uf2 is the default make BOARD=..., how can ...
manic glacierBOT
proven garnet
Can someone confirm I'm not accidentally lying to this fine human:
https://github.com/adafruit/Adafruit_CircuitPython_DS3231/issues/40#issuecomment-1006744828
That's been my experience from working with RTCs in CircuitPython so wanted to make sure
slender iron
@proven garnet I punted on it because I know it is complicated
proven garnet
Oh, gotcha, so a half truth then?
ember iris
Can someone confirm I'm not accidentally lying to this fine human:
https://gith...
The datetime library is full of holes and is really complicated and I absolutely feel for the original author(s) because I tried going through it to add to it and quickly realized how much of a pain 'time' is
slender iron
@proven garnet I think you are right. The library doesn't do it because it's complicated. An individual can do it for their needs when they need it.
One case is much more doable than all cases
proven garnet
I ended up doing one for a work project for DST, it was 
And that's just the USA's DST rules
slender iron
so you know better than I do π
proven garnet
Thanks for confirming!
half osprey
Hey! Just wanted to say that I watched this week's CircuitPython meeting on Youtube and it was so nice and pleasant, it reminded me of the weekly Mozilla meetings back when I worked for them. The hug reports warmed my heart! I hope I'll find some more things to contribute in the new year.
slender iron
Thanks @half osprey. Please reach out if you ever need guidance with contributing.
timber mango
Hello everyone, someone could help me, I have a QT Py RP2040, I need to send files to a USB memory stick. I found the USB_CDC Library and the USB_HID, which one can I use to transfer a file to the USB stick?
The QT is powered by 5 volts with a battery and if I put the USB stick, it should send the data to the USB stick.
stuck elbow
that's not possible, the rp2040 can't be an usb host
timber mango
the QT can be set as a Host
tulip sleet
CircuitPython doesn't currently support USB Host mode
timber mango
ahh π¦
tulip sleet
you can write to an SD card, would that serve your purpose?
timber mango
I imagine connecting a usb, can a micro sd be used?
tulip sleet
Example with a breakout (we have a number of boards that have uSD sockets): https://learn.adafruit.com/adafruit-microsd-spi-sdio
timber mango
as it is not possible, if it would be through the same connection that connects to the computer, where could I connect the usb.
In the QT Py RP2040 QT already has a USB connection, through which I connect to the MU, is it possible to use this connection?
tulip sleet
I"m not sure what you mean, do you want to send data to the host computer, or connect an SD card to USB? The former is possible, the latter not
#help-with-circuitpython is a more appropriate channel for this q
timber mango
I want to send a file from QTPyRP2040 to USB memory stick
timber mango
that is, at the moment it sends the data to the usb it is not connected to the computer.
tulip sleet
that is, at the moment it sends the data to the usb it is not connected to the c...
There are USB hosts and USB devices. It cannot act as a USB host right now
timber mango
what I want to do is to put data in the QT py RP2040, after a while I want to see the behavior of the sensor, i.e. the data I have stored. Now I want to access the data via a USB stick. What I wanted to do was to press a button and have the microcontroller pass the data to the USB stick. So I figured using the USB_CDC or USB_HID libraries, because as I look at the microcontroller matrix I can use them.
so in order to send data to a USB memory stick. So the usb memory stick has to be a host?
tulip sleet
so in order to send data to a USB memory stick. So the usb memory stick has to b...
USB devices connect to USB hosts. You cannot connect a USB device to another USB device if they are both acting as devices
so, no you cannot do that. You can write data to an SD card. Later you can read that data with a USB card reader
tulip sleet
@timber mango another more complete example: https://learn.adafruit.com/data-logging-with-feather-and-circuitpython/
timber mango
thak you, but is very big. i think it is better the first option, sd card,
I thought that the microcontroller is the host and a device was the usb memory stick.
devout jolt
Is there a runtime way to determine precision of floats? I know about sys.maxsize for ints. Is that a universal proxy for float precision?
tulip sleet
Is there a runtime way to determine precision of floats? I know about `sys.maxsi...
In CircuitPython, no. There is https://docs.python.org/3/library/sys.html#sys.float_info in CPython
You could try adding a small number to a larger number and see how small it has to be to not affect the result
devout jolt
You could try adding a small number to a larger number and see how small it has ...
I guess I'm also sort of asking "how is floating point precision specified in CircuitPython?" because I was assuming the "LONGINT_IMPL" Make define controlled both int size and float depth
tulip sleet
CPy floats are 30 bits, IEEE-754 32 bit floats with the lowest two mantissa bits dropped. H
They are not related to longints
devout jolt
ah, so it doesn't matter if it's a hardware FPU or software routines, the floats are always the same size across boards
tulip sleet
Except for the new RPi ports, maybe, but not sure about that
devout jolt
Right. okay that's very helpful. Thanks!
tulip sleet
Right. okay that's very helpful. Thanks!
This is helpful: https://www.h-schmidt.net/FloatConverter/IEEE754.html. Turn off the lowest two bits.
See https://github.com/adafruit/circuitpython/blob/main/py/mpconfig.h. We normally use MICROPY_OBJ_REPR_C.
daring ridge
Hoping this is an easy thing, so I'm asking here - if not, just let me know and I can open an issue.
My forked repo is a good bit out of date. So I'm updating locally - git fetch adafruit; get merge adafruit/main; git submodule sync --recursive; (and finally) git submodule update --init.
'git submodule update' (or with --init) fails with:
[bsd@yyz]:/circuitpython- git submodule update --init
fatal: 'bd34n' does not appear to be a git repository
fatal: Could not read from remote repository.
Please make sure you have the correct access rights
and the repository exists.
fatal: Fetched in submodule path 'ports/broadcom/firmware', but it did not contain bf96d0eda5952595d717fedb797aeb168483e9fa. Direct fetching of that commit failed.
fatal:
[bsd@yyz]:/circuitpython- git --version
git version 2.34.1
[bsd@yyz]:/circuitpython-
'bd34n' is my github id. Not sure why it thinks it's a repo. Not sure about the rest of the message - are special permissions needed for updating submodules? Any suggestions? Thanks!!
slender iron
Hoping this is an easy thing, so I'm asking here - if not, just let me know and ...
do make fetch-submodules at the top level
otherwise you may suck down gigs of broadcom firmware
daring ridge
That's working, thank you, @slender iron !
devout jolt
See https://github.com/adafruit/circuitpython/blob/main/py/mpconfig.h. We normal...
Oh I know why I was thinking they're connected. On SAMD21, time.monotonic() was less precise quickly as time went on, but that wasn't because of floating point precision but mapping of "ticks" to a float
tulip sleet
it was floating point precision ultimately.
devout jolt
yeah but not because of SAMD21 ahha
main furnace
Epsilon is the value you seek.
devout jolt
e.g. This prints out the same value on M0 and M4:
>>> import board; print(board.board_id, 1.2345678)
trinket_m0 1.23457
>>> import board; print(board.board_id, 1.2345678)
itsybitsy_m4_express 1.23457
yes, not the SAMD21 inherently. But the small boards don't have longints, so you can't use time.monotonic_ns()
manic glacierBOT
main furnace
epsilon = 0.01
while 1.0-epsilon < 1.0:
epsilon = epsilon / 2.0
print(epsilon)
What is the final value of epsilon?
devout jolt
epsilon = 0.01
while 1.0-epsilon < 1.0:
epsilon = epsilon / 2.0
print(epsilo...
Thanks! It turns out I wasn't really asking "what's the smallest floating point number representable?" but rather "Is there any difference in precision between different CircuitPython boards?" And the answer to that is "no"
main furnace
Ah, no problem.
devout jolt
It just took me a while to figure that out, ahaha
main furnace
Thanks for putting up with my math obsession. π
torn latch
CircuitPython doesn't currently support USB Host mode
This is extremely topical, I was just thinking about this myself. Context, I have a quite few 'abandoned' USB devices that I've used https://www.amazon.com/gp/product/B01EWW9R1E/ to rewrite drivers to become vanilla HID devices and useful once again. With the work in CircuitPython and HID, I was optimistic that I could update my spaghetti Arduino code. If USB Host ever did come to circuit python, I would certainly make use of it. Thanks Dan!
tulip sleet
This is extremely topical, I was just thinking about this myself. Context, I hav...
there is Arduino support: https://github.com/gdsports/USB_Host_Library_SAMD
not as easy, but it is a solution
slender iron
@devout jolt @main furnace I think it's useful to point out that this is due to the object representation we use: https://github.com/adafruit/circuitpython/blob/main/py/mpconfig.h#L79 Using other represetations could give us more precise floats at the cost of being allocated on the heap each time.
torn latch
there is Arduino support: <https://github.com/gdsports/USB_Host_Library_SAMD>
Yep, that's more or less what I'm using, the OG host shield library. just adding my 2c and a +1 that it'd be cool, but not required, in python
main furnace
I have a pi zero W around here somewhere...
slender iron
a non-W will work too
manic glacierBOT
manic glacierBOT
manic glacierBOT
Hi @tannewt
Where can I add "SOFTDEV_VERSION=7.0.1",
Can I add behind the "USB_PRODUCT = "Seeed XIAO nRF52840 Sense"
manic glacierBOT
proven garnet
Hey y'all! Does anyone have the HTS221 Temp & Humidity sensor? Looks like it's not sold anymore (or at least out of stock on Adafruit & Digi-Key)
If you do, I have a small request to run the simple test file on a PR branch of mine!
solar whale
If you do, I have a small request to run the simple test file on a PR branch of ...
I do have one, but it'll be a few hours before I can do any testing.
jaunty juniper
Hey y'all! Does anyone have the HTS221 Temp & Humidity sensor? Looks like it's...
I got it here on my desk
lone axle
@solar whale or @jaunty juniper if one of you end up having a moment. I think @proven garnet is likely referring to a quick test on this PR https://github.com/adafruit/Adafruit_CircuitPython_HTS221/pull/10. If you can check out that branch and try the simpletest in the examples dir using that branch to confirm it's working as expected that would be awesome.
jaunty juniper
Traceback (most recent call last):
File "code.py", line 10, in <module>
File "adafruit_hts221.py", line 188, in __init__
File "adafruit_hts221.py", line 267, in data_rate
File "adafruit_hts221.py", line 89, in is_valid
TypeError: can't convert 'int' object to str implicitly
let me check what that's all about
ember iris
```
Traceback (most recent call last):
File "code.py", line 10, in <module>
...
I've been digging into the stack trace too, that's a confusing one. (I don't have a sensor, I've just been stepping through the code)
jaunty juniper
(I am doing a review with fixes)
ember iris
Sweet! I was just about to test something but I'll wait for the fix. I'm not use to type hinting, and adding inheritance/setting attributes threw me for a loop. It'll be informative to see the fix
jaunty juniper
it's not related to types though, it's an unrelated change to the CV helper class
ember iris
I think that's where I was about to test. Types just make it slower for me to read while I'm new to it
jaunty juniper
there is a typing change I suggest, but I'm not sure: if a value can be 1 or 12.5 is it enough to type it as float or does it have to be Union[int,float] ?
ember iris
Not sure. I clearly misread the code pretty strongly, I thought the is_valid was using the rate label number (0 through 3) to make the check, not the rate name or the actual frequency of samples.
I think I need to walk away from this code for a bit. Do you know why the rate tuple is (str, index, rate)? With the change you suggested I can't find where or why the index would be used
onyx hinge
Oh I know why I was thinking they're connected. On SAMD21, `time.monotonic()` wa...
fwiw and you may already know this, but in 7.x we have supervisor.ticks_ms which wraps around in a specified way, and adafruit_ticks to wrap it (including the best compatibility with blinka & 6.3) so you can do "100ms from now" style deadline calculations without ever running into precision trouble, on the whole range of boards including those without long ints. (precision problems still exists on 6.3 without long ints, but we do our best)
devout jolt
fwiw and you may already know this, but in 7.x we have `supervisor.ticks_ms` whi...
Yes! I am very excited by supervisor.ticks_ms
proven garnet
Just checking now, thank you @jaunty juniper!!
jaunty juniper
I think I need to walk away from this code for a bit. Do you know why the rate t...
yeah it's not really useful here, it's the index of the entry in the list. I believe CV is a generic helper for enums, it dynamically creates a bunch of constants on the class, with the is_valid function to test if something is a valid value in the enum. So you dynamically define the possible values and simultaneously have a list of enums that can be used AND a list of valid values that can be tested.
and maybe the original does more, because it has that lsb property, but not in that library
proven garnet
Catching up now, yeah it seems like enum-like behavior, I've seen it in a few classes. This one doesn't use it to the full extent I've seen elsewhere hence why I wanted to reduce it
It's used commonly enough that I toyed with asking about turning it into a mini importable library
jaunty juniper
C python has some advanced structures helpers for structs, enums and such, this might be a pure-python adaptation of one of those
proven garnet
But it seemed antithetical to memory reduction for the basic functionality, that was my take
jaunty juniper
and it's so small, it would be kind of annoying to add one more dependency to a bunch of libraries
proven garnet
C python has some advanced structures helpers for structs, enums and such, this ...
That was my understanding
Alright, changes made per review
jaunty juniper
oh wait no I got it wrong
not in the code, but in my explanation
or understanding I should say
the values of the enums are the second value, the third value (renamed label) only serve as labels
proven garnet
Oh, yes that's correct
silent crane
Hello circuit python team. Have a question about the best way forward:
This line in the EPaperDisplay module is giving me a headache: https://github.com/adafruit/circuitpython/blob/afd686986ac3abf2fd8a7a060d8bf115de2bd8db/shared-module/displayio/EPaperDisplay.c#L148
Specifically I have a Waveshare 2.9" Epaper display and the LUT table that I have to send during start up is 154 bytes long. Because of the delay bit location I can't submit this table. Would it be best to subclass this with a different bit defined?
slender iron
what controller chip is it?
silent crane
I don't know as Waveshare refuses to publish that information. @stuck elbow gave me something to try before I go this route.
slender iron
what display is it? I've usually been able to find it
silent crane
Waveshare 2.9" V2
tulip sleet
this is under discussion right now on #help-with-circuitpython
silent crane
Yes, my apologies for crossposting, I had come to a bad conclusion.
slender iron
there is a different eink question in help-with
(it's easier)
I don't see a v2 here: https://www.waveshare.com/product/displays/e-paper/epaper-2.htm
slender iron
then I try to find the display here: https://www.good-display.com/product/6/#c_portalResProduct_list-15881346159377442-3
silent crane
I don't see a v2 here: <https://www.waveshare.com/product/displays/e-paper/epape...
Here is the wiki link. v2 replaced the original 2.9in display. https://www.waveshare.com/wiki/2.9inch_e-Paper_Module
ember iris
I don't see a v2 here: <https://www.waveshare.com/product/displays/e-paper/epape...
I think all of their screens are v2 by default as of around 5 ish years ago
slender iron
(I'm trying to find the datasheet so see what is says about the LUT command)
silent crane
I found the LUT table on Waveshare's RaspberryPi Python library. https://github.com/waveshare/e-Paper/blob/master/RaspberryPi_JetsonNano/python/lib/waveshare_epd/epd2in9_V2.py
lone sandalBOT
slender iron
@silent crane it looks like https://www.good-display.com/product/360.html and the SSD1680
it has the 0x32 command to send 153 bytes for LUT
(and the next command listed is 0x34)
orchid basinBOT
slender iron
I think the simplest thing is to add a kwarg for a two byte length field in the init list
and default it to false for the behavior we have now
digital shoreBOT
adafruit
winter mortar
32768 members soon!
slender iron
@solar whale you around to test more pi stuff?
solar whale
yes
slender iron
k, I'll make you a 2w build
2w
this has 10us delays after each EMMC register access
which the linux driver does afaict
solar whale
CanI just copy that to the boot partition of my SD Card?
I'm used to using Imager to load the whole.zip
slender iron
yup