(which means it has partitions, volumes, filesystems, etc.)
#circuitpython-dev
foggy cave
jaunty juniper
I guess you want to avoid writing to the same one twice ?
foggy cave
Does that mean one could use dd to copy the CIRCUITPY between two microcontrollers?
lone axle
Does that mean one could use `dd` to copy the CIRCUITPY between two microcontrol...
I think you can. But you may run into issues if anything else writes to either board and causes them to reset
jaunty juniper
but why ?
lone axle
Or actually I should say I think you could use cp I am less sure about dd
foggy cave
I think you can. But you may run into issues if anything else writes to either b...
Just unmount it :)
jaunty juniper
what reasons ?
foggy cave
what reasons ?
You're copying the entire partition, instead of dealing with the filesystem?
jaunty juniper
we are just copying files to a CIRCUITPY drive, why make it more complicated ?
jaunty juniper
don't play at the partition level, this is a microcontroller, not a hard rive
lone axle
I'm not really sure about their different functionalities or speed of either. But since most circuitpython devices are <10mb I think most any copying method will be pretty quick.
foggy cave
don't play at the partition level, this is a microcontroller, not a hard rive
But how does an OS mount a volume from USB without a partition?
Hm, I do have a MagTag or Funhouse or something, and I can try looking at Disk Management or diskpart and see
idle owl
I mean, I guess it's technically got a partition? But, for example, Disk Utility on Mac shows the microcontroller different from the other drives.
I don't know for sure, I'm simply relaying the info I know how to find.
foggy cave
I'm simply stating what I think I know
slender iron
the partition table of the CIRCUITPY drive is fixed
foggy cave
the partition table of the CIRCUITPY drive is fixed
No, I mean writing to the partition, not the entire disk
crimson ferry
best way to know is test it out on your device
i usually have multiple CircuitPython volumes, but rename them uniquely, so it’s easier on me, on editors, and for simultaneous copying
mental nexus
I've got the Devkit N8R8 board for the ESP32-S3. It looks like it flashed properly using the "UART" usb plug. When I switch over to the "USB" plug I see the CIRCUITPY drive. I'm not seeing a REPL (I tried both USB ports). Any tricks to getting a serial connection to the REPL on the ESP32-S3?
jaunty juniper
I'm getting the REPL on the USB port (and the ESP debug output on the UART port)
mental nexus
Interesting. Maybe I need to pull the latest from GitHub and rebuild. Alternately, Is there an easy way to flash from the UF2 files rather than build with the flash command?
jaunty juniper
I did install the UF2 bootloader
mental nexus
Can you point me to instructions on loading the UF2 bootloader?
Strange, just tried again and now I get REPL. Only thing I did was double click the reset button. weird.
jaunty juniper
I believe I took the devkitC one https://github.com/adafruit/tinyuf2/releases
and then it's gonna be erase_flash and flash combined.bin to address 0
mental nexus
Ok I’ll see if I can decipher the process. Thanks for the help!
jaunty juniper
I think I narrowed it down to that, though I remember some tinkering, maybe lowering the baudrate or something, or maybe it was with the C3 that it was flaky 🤷
esptool.py --port THEPORT --after=no_reset erase_flash
esptool.py --chip esp32s3 --port THEPORT --no-stub -b 921600 --after=no_reset write_flash 0x0000 combined.bin
proven garnet
This is good to know, I've encountered a few scenarios where Protocol seemed to be a good solution, particularly socket-like classes and things of that nature in CircuitPython
I didn't know much about __future__ or the annotations package within so thank you!
manic glacierBOT
solar whale
I am trying to use one of the V2 Feather ESP32 boards with MicroPython. According to the silkscreen, the I2C pins are SCL-20 and SDA-22 but when I try to use them in MicroPyhton i2c = I2C(0, scl=Pin(20), sda=Pin(22)) I get an "invalid pin" error ```MicroPython v1.18 on 2022-01-17; ESP32 module (spiram) with ESP32
Type "help()" for more information.
import head_13
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "head_13.py", line 24, in <module>
ValueError: invalid pin
stuck elbow
@solar whale maybe they are swapped?
solar whale
Am I using the correct Board def for this board?
solar whale
<@!343570430042308609> maybe they are swapped?
I'll try that
stuck elbow
also, might need I2C 1 instead of 0
solar whale
swapping them still gives invalid pin --- I'll try I2C1
still invalid on 1 .... 😦
the board says the module is an ESP32-PICO-MINI-02 with PSRAM so I think I have the correct build loaded...
solar whale
hmmm -- I works OK if I just move SCL to another pin -- I am using 14 -- I t will not work with scl on Pin 20
Wow - rshell is really nice with MicroPython --- This is the first time I have used it. Big step up from ampy....
stuck elbow
@onyx hinge I see the paralleldisplay now has a 16-bit mode, do you know if it is supposed to be functional on the esp32-s2?
because it hangs for me
onyx hinge
@stuck elbow at one point I had wanted to add 16 bit displays but I never got anywhere with it.
at least that's how I remember it going
stuck elbow
I see, that explains it
I'm trying to get https://www.makerfabs.com/wiki/index.php?title=ESP32-S2_Parallel_TFT_with_Touch(3.5'') working, but I can switch it to 8-bit with a jumper
lone axle
Anyone here using python 3.9 or 3.10? I'm curious if installing a library while specifying it's version is working in those environments:
pip install adafruit-circuitpython-lc709203f==2.2.1
Is it expected that this variable would be populated with a real value in the .py file that is downloaded from pip?
__version__ = "0.0.0-auto.0"
Or it's expected to stay as that 0.0.0-auto.0 value which is in the repo?
winter mortar
Anyone here using python 3.9 or 3.10? I'm curious if installing a library while ...
It worked for me with 3.10, with click==7.1.2
lone axle
It worked for me with 3.10, with `click==7.1.2`
Thank you for trying it out. Is click used by pip? I didn't know the version of that could make a difference. Are you able to share what OS you're on as well? Someone on github reported issues installing it on mac, I'm trying to narrow down what factors could lead to it if possible.
jaunty juniper
Is it expected that this variable would be populated with a real value in the .p...
it's populated by circuitpython-build-tools
when building the bundle (or each local release for each library)
lone axle
Ahh I see. So it does get populated into the copy that ends up in the bundle. But maybe not into the copy that gets sent to PyPi.
winter mortar
I'm on Mac. Re click, specifically what I was doing was overriding the version that requirements installed:
1064 pip3 install -r requirements-dev.txt
1065 pip install --upgrade click==7.1.2
I think that you are right. That matches what I am seeing. Just checked the file in the py bundle and one inside site-packages inside of a venv after installing.
jaunty juniper
❯ ag "0.0.0-auto.0"
lib/python3.9/site-packages/adafruit_platformdetect/chip.py
32:__version__ = "0.0.0-auto.0"
lib/python3.9/site-packages/adafruit_platformdetect/board.py
33:__version__ = "0.0.0-auto.0"
lib/python3.9/site-packages/adafruit_mprls.py
32:__version__ = "0.0.0-auto.0"
lib/python3.9/site-packages/adafruit_bus_device/spi_device.py
23:__version__ = "0.0.0-auto.0"
lib/python3.9/site-packages/adafruit_bus_device/i2c_device.py
21:__version__ = "0.0.0-auto.0"
lib/python3.9/site-packages/Adafruit_PureIO/spi.py
50:__version__ = "0.0.0-auto.0"
Ahhh, I understand now what you mean. Click was library you installed.
Are you willing to try this driver library?
adafruit-circuitpython-lc709203f==2.2.1
yeah installing with ==*** should work, this is not linked to __version__
winter mortar
Ahhh, I understand now what you mean. Click was library you installed.
That installed a lotta stuff, but included the right version
Successfully installed Adafruit-Blinka-7.1.0 Adafruit-PlatformDetect-3.20.1 Adafruit-PureIO-1.1.9 adafruit-circuitpython-busdevice-5.1.8 adafruit-circuitpython-lc709203f-2.2.1 adafruit-circuitpython-register-1.9.8 pyftdi-0.53.3 pyserial-3.5 pyusb-1.2.1
lone axle
I was thinking it was unrelated to that variable as well. The error reported here: https://github.com/adafruit/Adafruit_CircuitPython_LC709203F/issues/18 mentions differing versions expecting 2.2.1 but finding 0.0.0. The __version__ var was the only thing I could find with the zeros though. Even the METADATA file inside of site-packages is showing 2.2.1 for me. Not sure what could cause the issue Maybe some cached version in pypi infrastructure or something preventing it from actually downloading properly.
lone axle
Successfully installed Adafruit-Blinka-7.1.0 Adafruit-PlatformDetect-3.20.1 Adaf...
Thank you. Quite a perplexing issue. User was on Mac and python 3.10 as well though so good to know it's not causing trouble for everyone with that environement at least.
jaunty juniper
hmmm weird error
winter mortar
I was thinking it was unrelated to that variable as well. The error reported her...
The first two lines of the pip output for me were
Collecting adafruit-circuitpython-lc709203f==2.2.1
Downloading adafruit-circuitpython-lc709203f-2.2.1.tar.gz (28 kB)
In that issue, the first two lines are
Collecting adafruit-circuitpython-lc709203f==2.2.1
Using cached adafruit-circuitpython-lc709203f-2.2.1.tar.gz (28 kB)
Good catch! thank you
winter mortar
Looks like there was something funky about their cached version.
manic glacierBOT
manic glacierBOT
https://www.seeedstudio.com/Seeeduino-XIAO-Arduino-Microcontroller-SAMD21-Cortex-M0+-p-4426.html
Could some one make a PR? That would be great!
viscid pine
^ I'll leave rick some time to respond in case he wants to PR
manic glacierBOT
manic glacierBOT
proven garnet
There was an issue raised that Google returns the old RTD documenting site from before it got moved to circuitpython.org. Is there a way to close or redirect to the new site from the old one?
onyx hinge
The content should match on both URLs, as far as I understand it. So while we'd prefer people discover docs.circuitpython.org I don't think they're getting out of date information.
and https://circuitpython.readthedocs.io/projects/ov7670/en/latest/ [just a random choice] does redirect automatically to https://docs.circuitpython.org/projects/ov7670/en/latest/
is there an issue as in a github issue? I should review the details, if so.
lone axle
There was an issue raised that Google returns the old RTD documenting site from ...
I think maybe there is actually a secondary page that isn't up to date
Note that the issue links to https://adafruit-minimqtt.readthedocs.io/en/latest/api.html (no circuitpython in the URL)
This URL also exists: https://adafruit-circuitpython-minimqtt.readthedocs.io/en/latest/api.html and does redirect automatically to the new style docs URL / page.
@onyx hinge GH issue is here: https://github.com/adafruit/Adafruit_CircuitPython_MiniMQTT/issues/105
onyx hinge
ah that's yet another level of indirection away .. hmmm
lone axle
I don't know if https://adafruit-minimqtt.readthedocs.io/en/latest/api.html was meant to be an arduino docs page? The repos missing "CircuitPython" in their name tend to be arduino. But the content of the page appears to be for circuitpython and the download link on the left does link back to the circuitpython repo. But I'm not sure where this docs page is actually getting it's content from
onyx hinge
https://readthedocs.org/projects/adafruit-circuitpython-minimqtt/builds/ I wonder where the "copyright 2019" comes from; it has been re-built as recently as 5 days ago.
lone axle
This one would be for the one lacking circuitpython: https://readthedocs.org/projects/adafruit-minimqtt/builds/ it was built much longer ago only.
onyx hinge
so yeah there are a bunch of levels of indirection. Each 'project' gets its own blahblah.readthedocs.io site; then all the libs get placed under /projects/ of circuitpython.readthedocs.io, which now has the canonical location of docs.circuitpython.org.
lone axle
possibly all of the copyright dates on docs pages show 2019? at least here is another one that does. https://docs.circuitpython.org/projects/display_text/en/latest/
onyx hinge
docs/conf.py:copyright = "2019 Brent Rubell" yeah looks like the year is hard coded in the docs conf...
lone axle
Ah, my guess is that cookiecutter populates it from values given during the prompts. This one shows 2021 and my name: https://docs.circuitpython.org/projects/24lc32/en/latest/ I ran cookie cutter and made the initial commits in that library.
onyx hinge
In git, you can get the 'author date' of the latest commit with a command .. we could potentially run it during parsing of conf.py. git log --format=%ad --date=format:%Y -1
so that it would reflect the date of the last change built by RTD
(that'll print e.g., 2022, the 4-digit year)
ah, okay, so there are two similarly named things and the 'right' one does redirect and serve up the latest info but the wrong one doesn't. we could update the other one with a robots.txt that says never to index it and eventually that should make it disappear from search engine results.
solar whale
Has anyone tried using the new Feather ESP32 V2 with Micropython. The pin definitions for the SPIRAM build do not seem to be correct for the ESP32 PICO MINI 02 module on the board. Pin 20 cannot be used for I2C SCL (give an invalid pin) -- Is there some other build that will support the V2 or is it not supported yet.
I can "workaround" the problem by using a different pin for SCL (Pin 14 works fine) with the SPIRAM build
solar whale
hmm -- I just did a clean clone of the CP repository git clone https://github.com/adafruit/circuitpython.git then make fetch-submodules -- if if then do a git diff I get this discrepancy ```jerryneedell@Mac-mini circuitpython % git diff
diff --git a/ports/broadcom/firmware b/ports/broadcom/firmware
--- a/ports/broadcom/firmware
+++ b/ports/broadcom/firmware
@@ -1 +1 @@
-Subproject commit bf96d0eda5952595d717fedb797aeb168483e9fa
+Subproject commit bf96d0eda5952595d717fedb797aeb168483e9fa-dirty
If I cd into the ports/broadcom/firmware folder and do git diff I get
solar whale
interesting -- this issue #circuitpython-dev message only occurs on my Mac, not on my linux box...??
onyx hinge
<@&356864093652516868> Hey y'all, our weekly meeting is in about 30 minutes. Please add your notes to the document https://docs.google.com/document/d/1lQB-bN-XISGtWKES5DahLV3XOXG6elVJaedQBdsOd0Y/edit?usp=sharing and join us in the voice channel in a bit.
random junco
I can't make the meeting today, but i added a group hug - thanks to everyone who's listened and supported the podcast so far
solar whale
I can't make the meeting today, but i added a group hug - thanks to everyone who...
Listening to it right now! Well done!
random junco
thanks, appreciate it!
tulip sleet
we are finishing up an internal meeting, we'll be starting shortly
austere acorn
<@!196904073255976961> Please mute your mic 🙂
my bad, the computer had a brain freeze 😦
idle owl
my bad, the computer had a brain freeze 😦
No worries! You didn't make enough noise to come through, but I could see your name turn green in the list. Figured I'd have you catch it early.
lone axle
onyx hinge
We hope to get the recording and post it within the next week.
lone axle
solar whale
maybe there are so many that the signed int counter has rolled over 😉
stuck elbow
sounds plausible
idle owl
Organising Adafruit folks attending PyCon * 😄
onyx hinge
Congrats @proven garnet !
idle owl
🎉 @proven garnet
proven garnet
Thanks y'all! 
minor plume
Congrats, @proven garnet - that's exciting!
ember iris
That's fantastic to hear--congrats @proven garnet
onyx hinge
deshipu's projects: https://hackaday.io/project/184163-vegemite-sandwich-keyboard https://hackaday.io/project/178061-pewpew-s3
stuck elbow
🍢
idle owl
@solar whale Good to know! I'm supposed to get MicroPython going on the Feather ESP32 V2 for the guide.
solar whale
<@!343570430042308609> Good to know! I'm supposed to get MicroPython going on th...
more info here <#circuitpython-dev message>
hidden oxide
https://tenor.com/view/pusheen-cat-marshmallow-small-cute-gif-17307840
Yeah, I would too
No notes
slender iron
Congrats @minor plume ! 🎉
idle owl
Congrats @minor plume !
proven garnet
@minor plume 🥳🥳🥳
blissful pollen
Congrats @minor plume !
ember iris
Congrats!! That's Wonderful news @minor plume
hidden oxide
Congrats @minor plume
onyx hinge
@solar whale Have you considered filing an issue with micropython about the "I/O 21" issue? they are probably unaware of it.
solar whale
<@!343570430042308609> Have you considered filing an issue with micropython abou...
I thought about it, but held off since I was not sure I was looking at the correct board definitions. They only have a GENERIC_SPIRAM and I don't know what esp32 module it is targeted at.
there seem to be a few and it is not clear in MP how to distinguish them.
hidden oxide
mmm
slender iron
( I may have to leave in the next ten minutes or more because we have a person coming to quote us for a heat pump. They are supposed to arrive in 10 min.)
solar whale
Heat pumps are nice!! Enjoy!
slender iron
Ya, we're looking mainly to get the AC part
but not using gas to heat would be better too
and seattle is a good place to have a heat pump
solar whale
Thats what we thought, but use the heat more than expected
hidden oxide
there seem to be a few and it is not clear in MP how to distinguish them.
So MP doesn't have a adafruit_platformdetect.board function??
Or am I misunderstanding?
blissful pollen
All my long running wifi projects have a try block around the main routine. Not perfect but it does keep things going.
hidden oxide
I don't use MP that much
solar whale
So MP doesn't have a `adafruit_platformdetect.board` function??
not using Blinka
hidden oxide
Ahh, OK
idle owl
The only two WiFi projects I've done have try/except around the networking bits.
Like Mark said, not perfect, but it does keep going.
stuck elbow
I need to test this reload_on_error thing and see if I can scroll the error message on the PewPews with that
crimson ferry
perhaps some recourse on Safe Mode could help with more severe conditions
blissful pollen
I ran out of time when I was digging at the wifi code but I think looking at the code to make sure it doesn't get into any loops that never get returned. I saw something along those lines but it was buried deep.
ember iris
I've written a wifi manager class, that adjusts the try/excepts. The manager is suppose to make it easy to decide whether or not it retries right away, vs later. That's kept one of the esp32s2 feather online since.. November? Even when our router drops out
hidden oxide
https://www.tindie.com/products/deshipu/pewpew-lite-featherwing/ - these pewpews @stuck elbow ?
stuck elbow
https://www.tindie.com/products/deshipu/pewpew-lite-featherwing/ - these pewpews...
there is a whole family of those now, that was the first one
hidden oxide
Mmmkay....
blissful pollen
My outdoor (admittingly Arduino) project has the watchdog to give it more stability if wifi errors occur
stuck elbow
@hidden oxide the problem is we can't use displayio and repl with the 8x8 display, so I need to display the errors somehow myself
hidden oxide
So is this the repo? https://github.com/pewpew-game/pewpew
proven garnet
Haha no worries
stuck elbow
So is this the repo? https://github.com/pewpew-game/pewpew
More info: http://pewpew.rtfd.io/
hidden oxide
<@!746175873203437652> the problem is we can't use displayio and repl with the 8...
Ahh
solar whale
Life was so simple when editors just did text 😉
minor plume
Life was so simple when editors just did text 😉
I forget who said "the plain and simple truth is rarely plain and never simple". 🙂
hidden oxide
so then you need to display the errors outside of repl then??
stuck elbow
so then you need to display the errors outside of repl then??
right, there is already a menu program that runs as main.py and lets you select the game to run, what I was thinking, I could use the reload_on_error mechanism to get back to that menu on error, detect there has been an error and scroll it on the display
hidden oxide
Makes sense... esp if it sends you back to menu
proven garnet
Will stay tuned 
minor plume
Ditto. I'm happy to find other stuff to do and wait for the dust to settle, too.
hidden oxide
Boo DST
onyx hinge
canada and parts of mexico follow the same rule as the US, when DST is observed
idle owl
@proven garnet @minor plume Please continue as-is! The changes might come in how we do the imports, but we can always submit another PR at a later time.
solar whale
UTC -4, no?
onyx hinge
👋
minor plume
I live in AZ and most of AZ doesn't do daylight savings, but the whole world shifts around us, which is very confusing.
proven garnet
Thanks!
hidden oxide
Thanks!! cya next week!!
onyx hinge
@solar whale oh yeah I think you're right
minor plume
I live in AZ and most of AZ doesn't do daylight savings, but the whole world shi...
(For the curious, the part of Arizona that belongs to the Navajo Nation does do daylight savings, because part of their land is in AZ and part is in New Mexico, and they've decided they want their lands to be all in the same time zone.)
solar whale
That would be very helpful !
I thought I was the only one who had no idea what a "type hint" was....
idle owl
I thought I was the only one who had no idea what a "type hint" was....
Absolutely not!
ember iris
Thanks all!
solar whale
<@!343570430042308609> Have you considered filing an issue with micropython abou...
I did not want to "step on any toes" by creating an issue on a new board like this since as noted, I'm not sure what esp32 modules MP actually supports. I can certainly create an issue if you think I should. I posted this to the Adafruit Forum a few days ago, but have not had any response.
onyx hinge
@solar whale it doesn't seem like MP is going to be able to fix an issue if it's only discussed on Adafruit turf.
solar whale
<@!343570430042308609> it doesn't seem like MP is going to be able to fix an iss...
agreed -- I just did not know if the request should come from me of from someone at Adafruit.
I wanted to make sure Adafruit was aware of it since the product page says it works with MP
I considered a PR, but my MicroPython build environment is not working at present.
onyx hinge
fwiw looking at the code I don't see any way that any ESP32 (not S2 etc) could end up with that pin available, regardless of which "board", it's something at the "port" level. but that's not based on actual testing, soo
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/1I8mIuR4dXWFDHa32M216Mh_Nj655kqJh2MeRtj4GpLU/edit?usp=sharing
Take note: The US changes to Daylight Saving Time (“summer time”) between now and the next meeting, making the time 2PM in UTC-4. Use our online calendar with your favorite calendar app so you always see the meeting in your local time.
solar whale
fwiw looking at the code I don't see any way that any ESP32 (not S2 etc) could e...
It is not available on any of the other ESP32 modules that I have found -- only on the ESP32-PICO_MINI-02 used for the V2 board
slender iron
arm switched how they release their version of gcc: https://developer.arm.com/tools-and-software/open-source-software/developer-tools/gnu-toolchain
tulip sleet
arm switched how they release their version of gcc: <https://developer.arm.com/t...
the hardfloat version is separate: I wonder if they don't have M0 support?? A bit confusing
slender iron
that's a linux target, not bare metal
I'm gonna make a PR to switch. should free up room for deshipu's PR
I've been on 11.2 for a while
tulip sleet
what size diffs are you seeing/
slender iron
the matrix portal has a compilation issue. will let you know after I fix it
onyx hinge
"AArch32 bare-metal target (arm-none-eabi)" is the one we'd be using @slender iron ?
slender iron
yup
onyx hinge
for samd/etc
slender iron
though I can't get the pragma to actually work for the warning...
onyx hinge
If you are open source maker, then you can request a PID from http://pid.codes
Otherwise, companies should pay the USB-IF for a vendor ID: https://www.usb.org/getting-vendor-id
Oh nice, I didn't realize the message even directed what to do about the problem
idle owl
Do we submit board defs, or whatever MicroPython calls them, for Adafruit boards, or does that come from the community? There's no specific build for this Feather, and I'm not sure if there should be....
I'm now clearer on Jerry's fuzziness earlier.
There's not a lot of Adafruit boards listed, so I'm guessing it's not as common...?
onyx hinge
@idle owl no I don't think we do it on the regular
idle owl
I'm... not seeing easy to find info on how to add a board. There's "Porting CircuitPython", which I'm sure is wrong, and that says "A port will typically contain definitions for multiple boards" so I doubly don't think that's what I'm looking for. Probably this is not a path I should be heading down.
@solar whale What build did you use for the Feather ESP32 V2 to get MicroPython "going" on it? ("going" because I realise you ran into issues.)
solar whale
<@!343570430042308609> What build did you use for the Feather ESP32 V2 to get Mi...
GENERIC_SPIRAM -- I just used SCL on Pin 14
idle owl
Got it. Thanks.
solar whale
Do we submit board defs, or whatever MicroPython calls them, for Adafruit boards...
The confusing part for me is taht the board defs do not have a place for different pins. All the pins are defined for the general port -- It has a bunch of conditional compiles, but they are not really board specific.
idle owl
The confusing part for me is taht the board defs do not have a place for differe...
Really? I guess I assumed it was like CP. A forum post says there's a pins.csv file for each board...?
idle owl
Hmm.
manic glacierBOT
solar whale
hmm -- I may have missed something -- I did not see anything like that
solar whale
at least not for the esp32 port
idle owl
Wait... The UM Feather S2 has this: https://github.com/micropython/micropython/blob/master/ports/esp32/boards/UM_FEATHERS2/modules/feathers2.py
Wait is that a default file for it?
Or what's the deal there?
Ah helper lib.
solar whale
other ports are board specific https://github.com/micropython/micropython/tree/master/ports/samd/boards/ADAFRUIT_FEATHER_M0_EXPRESS
idle owl
Nevermind.
Hmmmmm
Why is ESP32 different I wonder? Because there's no remapping done?
onyx hinge
@slender iron So I don't know if it's a right fix but I added CFLAGS += -fno-builtin-memcpy and matrixportal builds.
slender iron
How much space is left?
onyx hinge
I don't know WHY gcc thinks that memory region starting at QSPI_AHB is only 0 bytes long (I can't see any reason why it should) but it stops concluding that the improper write is occurring when it is told not to treat memcpy as working like the standard function of that name
||
|| 494976 bytes used, 4736 bytes free in flash firmware space out of 499712 bytes (488.0kB).
Ja was the one failing
onyx hinge
trying it now
maybe i'm just impatient but the builds feel kinda slow.
500764 bytes used, -1052 bytes free in flash firmware space out of 499712 bytes (488.0kB). drat
slender iron
Bummer. I’m off for lunch and walk
jaunty juniper
interesting -- this issue https://discord.com/channels/327254708534116352/327298...
that's a filesystem not case sensitive thing, they have two files with identical names only the case differs. If you didn't setup your mac drive in case sensitive, it will interpret 2 file entries as the same file and one will be considered wrong
so like it should have xt_RATEEST.ko and xt_rateest.ko but it only has the latter, so the former is marked as "modified"
❯ git st .
## HEAD (no branch)
M xt_DSCP.ko
M xt_HL.ko
M xt_RATEEST.ko
M xt_TCPMSS.ko
❯ ls | grep -i xt_RATEEST.ko
xt_rateest.ko
Interesting, this is the first time I’ve run into it. Thanks.
jaunty juniper
if you ask me, having 2 files with the same name that only differs by the case is not user friendly, we humans don't "see" case that much when we think about file names, but there it is
solar whale
It’s good to know what happened.
jaunty juniper
I don't build for broadcom, so I just ignore it, though on APFS it would be rather easy to add a small case-sensitive partition just for that, rather than reformat or use a disk image
actually it doesn't have to be small, they share space anyway
solar whale
I have one I used to use. But I’ll just stay on my Linux box.
timber mango
Just out of curiosity, how hard would it be to add circuitpython support for a custom board I was making?
I have a project I want to make that is based on a SAM E51 (Adafruit already has support via the Feather M4 CAN Express) that I would like users to have the option of using Arduino and circuitpython on. The difference between my board and the Feather board is that my board will have specific headers broken out for specific applications
stuck elbow
Just out of curiosity, how hard would it be to add circuitpython support for a c...
take a look at the ports/atmel-samd/boards directory, each directory is one supported board
slender iron
@timber mango the trickiest thing can be adding support for a new external flash chip
@onyx hinge how did you add that flag? I tried $(BUILD)/ports/atmel-samd/supervisor/qspi_flash.o: CFLAGS += -fno-builtin-memcpy and it still failed to compile
onyx hinge
<@!426200547024961539> how did you add that flag? I tried `$(BUILD)/ports/atmel-...
diff --git a/ports/atmel-samd/Makefile b/ports/atmel-samd/Makefile
index b9bc36344..b6d8b75b0 100644
--- a/ports/atmel-samd/Makefile
+++ b/ports/atmel-samd/Makefile
@@ -158,7 +158,7 @@ else
endif
endif
-CFLAGS += $(INC) -Wall -Werror -std=gnu11 -nostdlib -fshort-enums $(BASE_CFLAGS) $(CFLAGS_MOD) $(COPT) -Werror=missing-prototypes
+CFLAGS += $(INC) -Wall -Werror -std=gnu11 -nostdlib -fshort-enums $(BASE_CFLAGS) $(CFLAGS_MOD) $(COPT) -Werror=missing-prototypes -fno-builtin-memcpy
ifeq ($(CHIP_FAMILY), samd21)
CFLAGS += \
I applied it to all files
slender iron
the error is coming from linking which is why global cflags works
I disabled the errors instead
my build is -392 even though it has the -Os lib
slender iron
The 11.2 doesn't look like the libs were stripped so it's huge
I tried stripping it but apparently did too much
may look tomorrow
tulip sleet
i'm looking at their bugtracker
forum more active: https://community.arm.com/support-forums/f/compilers-and-libraries-forum/?
timber mango
<@456226577798135808> the trickiest thing can be adding support for a new extern...
What if I use a flash chip that’s already supported? Is there a list somewhere?
solar whale
@idle owl @tulip sleet FYI -- take a look at this discussion <#circuitpython-dev message> -- it looks like the case sensitive file system is once again necessary for the broadcom builds...We discussed this awhile ago and found it was not needed for the espressif builds any more, but ... it's back -- I just verified that it woks fine with a case sensitive file system but not with the "standard" MacOS file system. sigh,,, I guess I'll go back to using my case sensitive partition when I want to build on the Mac.
solar whale
ugh -- I guess i never tried a broadcom build on my Mac . Now I ham having issues getting the correct version of gcc for the M1 Mac -- I think I'll stick to Linux for the broadcom builds -- probably for all builds...
solar whale
There appears to be an issue with the arm compiler (aarch64) and rosetta on the m1 mac. I suppose I could try building gcc locally, but for me, it is not worth the trouble to resolve just for the broadcom ports.
tulip sleet
Just out of curiosity, how hard would it be to add circuitpython support for a c...
We have a guide: https://learn.adafruit.com/how-to-add-a-new-board-to-circuitpython
tulip sleet
@onyx hinge if you are willing to approve #6117 (from future) then I can merge #6118 immediately after that. tnx
onyx hinge
@tulip sleet I'll take a look right now
tulip sleet
ty
slender iron
black (formatter) failed
it reformatted adafruit_vl53l0x.py
ya, we have standard formatting
stylistic
vscode probably has a black extension you can install
it's a standard python formatter
🙂
timber mango
We have a guide: <https://learn.adafruit.com/how-to-add-a-new-board-to-circuitpy...
Thank you and thank everyone who pointed me in the right direction. Some developer communities are not as friendly
idle owl
<@!330227457296957440> <@!329766224093249548> FYI -- take a look at this discus...
I kept my CP build environment in a case sensitive image anyway. So that works out for me.
viscid pine
@tulip sleet should i open a pull request to 7.2.x or is the backport note for yourself?
tulip sleet
<@!329766224093249548> should i open a pull request to 7.2.x or is the backport ...
a PR would be great - you can just cherry-pick the commit(s). thanks!
viscid pine
🍒
orchid basinBOT
slender iron
@tidal kiln mind taking a look at this when you have a chance? https://forums.adafruit.com/viewtopic.php?f=60&t=189077&p=915505#p915505
jaunty juniper
I get that, it's very noisy
slender iron
thanks! we'll need an issue to fix it
manic glacierBOT
onyx hinge
@lone axle you were asking the other day about the 'version' marker in __init__.py. I think that it is SUPPOSED to be correct in pypi. It's these two lines that are supposed to do it:
use_scm_version=True,
setup_requires=["setuptools_scm"],
https://github.com/adafruit/Adafruit_CircuitPython_LC709203F/blob/main/setup.py#L27
Maybe something's become broken about it? Or maybe I'm mistaken.
I see it's become different in current cookiecutter, maybe it just needs to be updated? https://github.com/adafruit/cookiecutter-adafruit-circuitpython/blob/main/{{ cookiecutter and 'tmp_repo' }}/{%25 if cookiecutter.pypi_release in ['y'%2C 'yes'] %25}setup.py{%25 endif %25}#L60
tidal kiln
<@!337989520962093057> mind taking a look at this when you have a chance? <https...
yep. noisy for me also. looks like this might be a soon-to-be-created core issue. need me to test anything else?
lone sandalBOT
lone sandalBOT
slender iron
yep. noisy for me also. looks like this might be a soon-to-be-created core issue...
Nope. Thanks! I'll make an issue and then post a link to the forum
orchid basinBOT
idle wharf
<@!382939733107408897> you were asking the other day about the 'version' marker ...
and @lone axle here's when @viscid pine added that SCM Version change he could likely advise on your libraries implementation of same.
https://github.com/adafruit/cookiecutter-adafruit-circuitpython/pull/146
viscid pine
wrong @ ? i dont know what this is
slender iron
wrong purple person 🙂
@onyx hinge did you do anything with the touchscreen on the s3 box?
onyx hinge
no I don't think I got that far
slender iron
kk, you got pretty far! it's nice that the screen comes up
viscid pine
I got one of these today: https://github.com/Xinyuan-LilyGO/T-01C3, is it ok to add it to circuitpy?
it doesn't expose any usb
slender iron
yup, that's ok for C3. we'll have ble at some point
viscid pine
it's kind of cool, it's a c3 replacement for the esp8266 boards with the same pinout
lots of compatible carriers that it should work with
onyx hinge
@slender iron I don't think anything in particular stopped me from working on the touchscreen unless it was an i2c problem, is that how it's interfaced? I think i2c is better now
slender iron
ya, it is i2c. datasheet is under nda though 😕
but there is this source
that should be enough
slender iron
@mental nexus may have looked into touchscreens too
idle wharf
wrong @ ? i dont know what this is
Sorry @viscid pine ..
mental nexus
<@!366059101203202052> may have looked into touchscreens too
I never studied the S3 Box since none were available for a while and I never found final specs for the parts it was using other than that the display is SPI (I think). For touchscreens recently I extended the FocalTouch library for a chip from the same supplier.
slender iron
cool cool. it looks like that driver will generate a bunch of dicts so it might be nice to have a non-alloc version in the future
I'll copy it for now
it looks like the touch chip is a tt21100
lone axle
I see it's become different in current cookiecutter, maybe it just needs to be u...
Ah, yes I think this is it. The libraries that have been made after that change to cookiecutter do end up having the correct version number in them, I just tested it with one of the newer ones. Thank you to @idle wharf also. I think we could update the others with a sweep in the libraries if we want to get the versions embedding into them when installed via pip.
tough flax
Check this out @slender iron and @tulip sleet
Got the boards today
One more
I have this one too bu I’m going to rework it to have a male jack instead of the pcb jack
tulip sleet
👍 👍
tough flax
Thanks for all the work on the usb host stuff.
I need to head home and watch all the streams
tulip sleet
scott is the one doing all of it (and Thach)
tough flax
Thank you @gentle bronze !
I love the way the host #bff works with only the one set of headers. So slim and tiny!
Easier to see with my hand for perspective
manic glacierBOT
slender iron
@tulip sleet what is your release plan? I think it'd be good to move to 7.3 after 7.2.1 is out
tulip sleet
if you feel the 1.18 merge is stable, etc., I wonder if I should even bother with 7.2.1.
slender iron
ya, I'd skip alpha
we haven't had any pre-releases on 1.18
so we should definitely do one
but I haven't personally had any issues
7.2.1 is basically ready after these backports afaict
so we might as well do it
tulip sleet
https://github.com/adafruit/circuitpython/issues/6115 was in 7.2.0 but I'll move it forward
tulip sleet
@slender iron should I make a 7.2.1-rc.0 or just do it? Main fix from my point of view is the auto-reload stuff
tulip sleet
then I'll make release notes and release it by tonight
slender iron
thank you
slender iron
@tough flax did you start the intellikeys pyusb code?
tough flax
No, I have Ella coming down for a visit and have been swamped... I got the PCBs and snuck in time for assembly 🙂
Very happy with the hardware though
I am thinking about installing the C-API version that you referenced just to see if the hardware works
slender iron
I am thinking about installing the C-API version that you referenced just to see...
Thach's WIP is here: https://github.com/hathach/tinyusb/commits/pio-host
I've distracted him a bit from it requesting a lower level host api
his work for that is in the more-host-bare-api branch
idle owl
@solar whale https://github.com/micropython/micropython/pull/8406 🙂
solar whale
<@!343570430042308609> https://github.com/micropython/micropython/pull/8406 🙂
Woohoo! Thank you!
idle owl
You're welcome! First contribution to MicroPython 😊
tough flax
Thanks @slender iron
slender iron
@lone axle I've got the start of a painter on touchscreen now
pleasantly responsive too
lone axle
<@!382939733107408897> I've got the start of a painter on touchscreen now
I think this is the one I saw: https://github.com/adafruit/Adafruit_Learning_System_Guides/tree/main/CircuitPython_PyPaint
slender iron
thank you!
ok, I won't go crazy with mine. that looks neat
could be upgraded to bitmaptools
lone axle
It could use an adjustment for a maximum size maybe as well. PyPortal Titano has memory allocation error running it
tulip sleet
@jaunty juniper another WSGI server: https://github.com/deckerego/ampule
tulip sleet
@slender iron @onyx hinge fyi I found a regression in the autore-load fix in the proposed 7.2.1: it will not auto-reload during fake deep sleep. Fixing that, but that seems to reintroduce some of the other older problems. So no 7.2.1 tonight. I'll wait until next week.
onyx hinge
@tulip sleet I'm glad you found that before you shipped it. waiting a bit longer can't hurt.
mental nexus
it looks like the touch chip is a tt21100
Can’t find much detail on this touch controller chip TT21100. Looks like it was made by Cypress but may have been discontinued https://www.mouser.com/PCN/Cypress_Semiconductor_PTN181204.pdf. Maybe now made by Parade Tech, but I can’t find a source with it available. Perhaps this product uses some NOS stock displays? Interestingly your new repo is actually one of the top hits on the big search engine.
manic glacierBOT
slender iron
@tulip sleet where do you think the utf16 file should go?
tulip sleet
<@!329766224093249548> where do you think the utf16 file should go?
I think Damien created the shared directory for that purpose
some but not all are submodules?
tulip sleet
the README states the purpose
i wasn't sure about readline, etc. but I guess those are copies
could be in shared/runtime
slender iron
I have it in shared-module/usb/utf16le.* for now
I hesitate to put it in an existing subfolder because it'd be better if they matched mp
tulip sleet
probably it will only get used in USB; I think there is also some UTF16 code in the USB device layer, but maybe it's in tinyusb
I remember writing something, but maybe it was in python for descriptor strings
slender iron
I offered to put it in tinyusb and thach didn't want it there
the outer function is mp specific anyway
tulip sleet
it could be a long-term thing, not a big deal
slender iron
I borrowed a bunch of the code from FATFS anyway
ok, I've at least split it apart now
manic glacierBOT
tidal kiln
@lone axle ping. non-urgent CP text question.
lone axle
<@!382939733107408897> ping. non-urgent CP text question.
I got some time, what's up?
tidal kiln
hey. i seem to remember someone putting together some basic UI elements. like a text list, etc.?
essentially extensions of label, as i recall
lone axle
hey. i seem to remember someone putting together some basic UI elements. like a ...
There is this library: https://github.com/adafruit/Adafruit_CircuitPython_Simple_Text_Display
that one is like a wrapper around Label that takes care of some of the more specific details around text.
tidal kiln
cool. thanks. that might be it. let me take a look.
lone axle
I've also used GridLayout with a single column and multiple rows in the past to create a vertical list of text items. https://github.com/adafruit/Adafruit_CircuitPython_DisplayIO_Layout/blob/main/adafruit_displayio_layout/layouts/grid_layout.py I used it as menu that the user scrolls through in this project: https://github.com/FoamyGuy/CircuitPython_Blink_A_Sketch/blob/f8e0dd7e8e70e5e7cdcdd6a6e8a3fc7666194ed2/code.py#L80-L87
Kmatch also created this one recently: https://github.com/circuitpython/DisplayIO_ScrollBox it's a scrollable text box so it can hold more text than will fit visually in the rectangle it occupies and be scrolled up and down by the user.
lime trellis
Is there anything special to get CP to build with CIRCUITPY_UHEAP on the atmel-samd port?
just enabling it in the board definition causes conflicts with gc
tidal kiln
@lone axle sry. got distracted. thanks for the links! those look like what i was trying to remember.
old smelt
Thanks @gilded cradle . I saw this and then immediately forgot to comment. Anything you still need from me?
gilded cradle
Thanks <@!478616919721771019> . I saw this and then immediately forgot to comme...
No worries. If everything looks good to you, then no changes are needed.
slender iron
Is there anything special to get CP to build with CIRCUITPY_UHEAP on the atmel-s...
may just be outdated. its not on in any build by default
should probably be removed
viscid pine
golfing a minimal asyncio example a bit:
import asyncio
async def printy(name, period):
while True:
print(name)
await asyncio.sleep(period)
asyncio.run(asyncio.gather(printy("t1", 0.7), printy("t2", 1)))
gather automatically promotes to task, no need to create_task, gather returns an awaitable, pass that straight to run instead of creating a main func. Edit: turns out it isnt compatible with cpython
manic glacierBOT
lime trellis
may just be outdated. its not on in any build by default
is there anyway to query the heap size of an object without uheap?
manic glacierBOT
mental nexus
I’m just starting to evaluate the ESP32-S3 for driving RGB LCD displays. I’ve hacked the lcd RGB esp-IDF example code to strip out the LVGL parts and just draw a colored square “bouncing” along the screen.
I’m struggling to understand how all the timing works for defining each frame.
In my code, I use esp_lcd_panel_draw_bitmap to draw my colored 50x50 pixel box along its motion path inside a while True loop. What appears to be happening is that each frame draws just one new position movement of the box (move by x and y by 1 pixel each). It seems like I can only get one draw_bitmap call in each frame. That doesn’t seem right to me, I suspect something must be blocking to prevent multiple draws in a frame.
I expected that this box would fly around the screen since the while loop would be updating much faster than the frame rate. But alas it’s not the case. Any suggestions on how to understand the sequence of events of queuing up several calls to draw_bitmap and have them all happen before the next frame?
Frame rate is about 12 fps on an 800x480 display, it’s a start.
There is a call back function in the display configuration but I’m not exactly sure how to use it.
.on_frame_trans_done
Can I do a bunch of bitmap drawing and then trigger the next frame? Do I have to do all the work in the callback routine?
manic glacierBOT
mental nexus
There is a call back function in the display configuration but I’m not exactly s...
Digging into the esp-IDF source code a bit more, I’m beginning to suspect that the draw_bitmap code triggers a new frame update. https://github.com/espressif/esp-idf/blob/a8ba5a0264d50f9aadd79838e2672a2828704770/components/esp_lcd/src/esp_lcd_rgb_panel.c#L337. I guess I’ll need to brush up on semaphores to understand how to hack this code to update several different regions before triggering a frame update.
onyx hinge
https://nedbatchelder.com/blog/202203/dinghy_digests.html neat looking tool, should try it on circuitpython related repos
jaunty juniper
is this worth addressing ?
>>> rainbowio.colorwheel(-10)
-7680
tulip sleet
is this worth addressing ?
```py
>>> rainbowio.colorwheel(-10)
-7680
```
I think the original Python impl didn't check for negative either.
jaunty juniper
pypixelbuff and Blinka do:
if pos < 0 or pos > 255:
return 0, 0, 0
it would be fine to fix it. the >255 check is already a recent fix.
jaunty juniper
what would be the preferred fix ? black below 0 and % 256 above ? (add return 0 if < 0 in builtin rainbowio)
onyx hinge
Personally I wish it worked for any number like sin.. though if it runs into trouble after 224 or 230 that's less ideal and it's "good" to cause the user to think about what happens after 255.
jaunty juniper
I would prefer just using % 256 which works great in python for negative numbers too, but how much more complex is it for the C code, that had to be trimmed down to fit ?
viscid pine
&255?
onyx hinge
That work s for negative numbers, good call
tulip sleet
I would prefer just using `% 256` which works great in python for negative numbe...
Do you want to change pypixelbuf and Blinka too?
should they be more consistent (and is % more useful?)
jaunty juniper
yeah, I noticed the discrepancy when posting code for blinka in the forum and was about to submit a PR to blinka when I noticed the negative numbers issue in the builtin module
I like the time trick for a rainbow animation for its simplicity but had to add a % 256 on blinka
colorwheel( time.monotonic() * some_speed_factor )
tulip sleet
I would say we could modify blinka and pypixelbuf to do the modulus. It's more useful. I doubt if any example relies on out-of-range going to (0,0,0)
manic glacierBOT
orchid basinBOT
proven garnet
Does displayio.FourWire.send() accept a str as the data argument? It's typed as circuitpython_typing.ReadableBuffer but I saw recently that str happens to implement that protocol but is not behavior that is intended. Should the argument be typed as a Union or should libraries (I think I see one) that use a str (it's typically a blank string) be changed to b""?
stuck elbow
I think it doesn't care internally, it just takes the bytes from the memory
you could also use an array
tulip sleet
Does `displayio.FourWire.send()` accept a `str` as the `data` argument? It's ty...
Make it Union[ReadableBuffer, str] since that's the most correct. The type-checking tools are not going to know that str gets treated as a bunch of bytes.
since that's not true for CPython (regular Python)
proven garnet
Make it `Union[ReadableBuffer, str]` since that's the most correct. The type-che...
Thanks, I'll submit that fix!
onyx hinge
Make it `Union[ReadableBuffer, str]` since that's the most correct. The type-che...
@proven garnet @tulip sleet The other position to take would be that the type annotation gives the "ideal" types, even if a relaxed type check within CircuitPython means that specifying some other type today (str) is not rejected. It's not "here's all the things that in fact work today", it's "here's the contract we make about what types will work now and in the future; and what types need to work in an alternate implementation, such as blinka".
tulip sleet
the previous case of this were SSID and password in wifi.radio, and there it seems quite fine to take str's. For displayio.FourWire.send(), it makes sense to only take buffer protocol, since you really want bytes. But if we don't change our examples not to use str, or validate the argument not to use str, then there's a case where it works, but not as documented (too permissive).
lone axle
Good day to all <@&356864093652516868> 👋 The weekly meeting will begin in about 1 hour and 10 minutes in the CircuitPython voice channel here on Discord. If you're participating please add your notes to the document: https://docs.google.com/document/d/1I8mIuR4dXWFDHa32M216Mh_Nj655kqJh2MeRtj4GpLU/edit?usp=sharing
onyx hinge
Surprised by the meeting time shift? It's because of "summer time" observation in most of the US. View our calendar online in your timezone: https://open-web-calendar.herokuapp.com/calendar.html?url=https%3A%2F%2Fraw.githubusercontent.com%2Fadafruit%2Fadafruit-circuitpython-weekly-meeting%2Fmain%2Fmeeting.ical&title=CircuitPython Meeting Schedule&tab=agenda&tabs=month&tabs=agenda or add it to your calendar app https://raw.githubusercontent.com/adafruit/adafruit-circuitpython-weekly-meeting/main/meeting.ical
proven garnet
the previous case of this were SSID and password in wifi.radio, and there it see...
@tulip sleet @onyx hinge those are really good points. I think mostly what would worry me is that it isn't just examples but other libraries relying on str being acceptable, and what would cascade if that ever changed for some reason. I guess if it really does rely on the Buffer protocol, maybe it just makes sense to make sure we're using something that will always rely on that.
I don't know much about the core so I don't know what the likelihood or reason that behavior would change, but if it would, maybe it makes sense to futureproof it?
(fix libraries using str as opposed to fixing typing to include str that happens to work right now)
tulip sleet
<@329766224093249548> <@426200547024961539> those are really good points. I thi...
also @onyx hinge another consideration is whether we should actually prevent str's from being buffer protocol in some or all cases
that would be an incompatibility with micropython
we need it internally because there is a lot of shared code betwen strs and bytes, but we could prevent memoryview from taking a str, etc.
stuck elbow
what would be the advantage of doing that, in practical terms?
tulip sleet
only preventing incompatibility with CPython
e.g. memoryview(str) in CPy/MPy provides access to the raw UTF-8 bytes, not possible in CPython
idle owl
@lone axle Are you still available after the CP Weekly to work with Eva and I for her to teach us about Adabot patches?
lone axle
<@!382939733107408897> Are you still available after the CP Weekly to work with ...
Yep, I am.
idle owl
Yep, I am.
Ok, excellent. We'll start after you do the post-meeting whatnot. So let me know when you're done.
proven garnet
e.g. memoryview(str) in CPy/MPy provides access to the raw UTF-8 bytes, not poss...
I'm happy to file issues with the offending libraries if we decide to go that way. Should be quick fixes too, should just be changing "" to b"" from a quick search on GitHub.
onyx hinge
Added!
idle owl
Thanks Jeff!
onyx hinge
@candid sun you might have to close out and re-open the voice channel with the added role
maybe do a mic check before recording starts
candid sun
maybe do a mic check before recording starts
thank you!
stuck elbow
lurking
turbid radish
lurking
stuck elbow
meow
blissful pollen
lurking today, didn't have any updates
mental nexus
🐕🦺 🤷
turbid radish
I have a dog too
stuck elbow
leave cat and dog notifications enabled
turbid radish
Yay Paul!
random junco
thanks. 🙂
minor plume
The Apoca-pi Now is proof that anything worth building is worth over-building. 🙂
turbid radish
Tag on Twitter & email please
idle owl
🎉
candid sun
yay! glad it was helpful 🙂
blissful pollen
Good problem: always too many things to want to learn and explore.
Great solution: lots of guides help you get up to speed fast!
slender iron
🎉
idle owl
🎊
turbid radish
yay liz
blissful pollen
Congrats @candid sun ! Super exciting
candid sun
thank you all!
minor plume
Congrats @candid sun !
candid sun
that would be awesome @onyx hinge
stuck elbow
great to see you here @candid sun!
random junco
Joining Adafruit on Pi Day will make it easy to remember you work anniversary. 🙂
idle owl
Joining Adafruit on Pi Day will make it easy to remember you work anniversary. ...
That came up in our meeting before this 😄
onyx hinge
And if you're going to be at pycon, please feel free to open a DM or otherwise chat with me if you are interested in meeting in the hallway track. I will be there the conference days but will leave as the sprints get started.
idle owl
Melissa and I will be attending the day before the conference (Education Summit), and through the Sprints. Feel free to message me as well if you'll be attending PyCon 2022!
onyx hinge
@idle owl wow what a coincidence. want to get together while we're there?
turbid radish
Thanks @minor plume for your contribution to the newsletter and for broadcasting your making
idle owl
<@!330227457296957440> wow what a coincidence. want to get together while we're ...
Totally! 😄
random junco
I hope to cross PyCon off my bucket list next year
blissful pollen
I hope to cross PyCon off my bucket list next year
I've heard of it and kinda wish I had more time this year to figure out how I could attend. I miss conferences
minor plume
I hope to be able to attend PyCon next year too...but it's hard to think that far into the future when the world still feels so uncertain right now.
onyx hinge
I've been to one PyCon (2002? a looong time ago, I think it was the inagural one) and one PyOhio (2019). Good experiences.
hope to make pycon a more regular feature of my life while i'm working on CircuitPython, which I have no plans to stop
stuck elbow
travelling to events feels so unreal now
minor plume
travelling to events feels so unreal now
Totally does.
stuck elbow
i will be running circuitpython workshops on EuroPython this year
random junco
when is EuroPython?
random junco
Thanks
stuck elbow
https://ep2021.europython.eu/
that's last year
onyx hinge
@mental nexus curious how many/any I/Os are left after connecting to the lcd & touchscreen
idle owl
Congrats @mental nexus!
onyx hinge
hah just noticed I pasted 2021 info 
blissful pollen
Getting your first PCB is a real awesome feeling @mental nexus good job!
random junco
I could have hit Google, too. 🙂
mental nexus
<@!366059101203202052> curious how many/any I/Os are left after connecting to th...
Right now I’ve just got six left (and keeping several not in use that I think are required left alone for PSRAM pins 35-39).
And plus I’ve left the USB pins free (19 and 20)
mental nexus
Oops and one in my six is the Boot button so I think one must be cautious with that one.
mental nexus
Getting your first PCB is a real awesome feeling <@!366059101203202052> good job...
Fingers crossed! Amazing that PCBs can be made and populated in small volume in just a few days and “palatable” cost for hobbyists.
jaunty juniper
I was surprised of the big changes made to remove so-called "marketing statements"
onyx hinge
I suspect micropython deliberately made trade offs that are different than ours, but I wouldn't call it "lazy".
🥧
random junco
@idle owl When is CircuitPython day again in 2022? I can't find my notes (sorry!)
turbid radish
please subscribe!!
idle owl
<@!330227457296957440> When is CircuitPython day again in 2022? I can't find my...
August 19
turbid radish
Tell your friends and others to subscribe too
onyx hinge
@turbid radish can we make subscriptions for our cats and dogs, or is it just for people?
random junco
thanks. I just noticed earlier today that the 5 year anniversary of the 1.0 release is this July
onyx hinge
thanks tim!
mental nexus
Thanks all!👋
random junco
thanks foamyguy!
gilded cradle
Thanks everyone
candid sun
thanks everyone! and great job hosting @lone axle !
turbid radish
Thanks
mental nexus
Anyone with an email address
tidal kiln
Adafruit CircuitPython 7.2.1 on 2022-03-14; Adafruit Matrix Portal M4 with samd51j19
>>> from adafruit_ble import BLERadio
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "adafruit_ble/__init__.py", line 21, in <module>
ImportError: no module named '_bleio'
>>>
am i missing something?
slender iron
looks like it was disabled: https://github.com/adafruit/circuitpython/commit/8f741e77675e0b773579fe1cf2384e446ea4fb5e
tidal kiln
@slender iron thanks.
Adafruit CircuitPython 7.1.0 on 2021-12-28; Adafruit Matrix Portal M4 with samd51j19
>>> import _bleio
>>>
does it need to be put back in?
slender iron
yup
tidal kiln
more general question. since this:
https://learn.adafruit.com/adafruit-airlift-breakout/circuitpython-ble
could be on potentially any board, is there any case where it should not be included?
idle owl
This reload thing is awful. This over and over while copying libs to the board. Ended with it working, apparently.Auto-reload is on. Simply save files over USB to run them or enter REPL to disable. code.py output: Traceback (most recent call last): File "code.py", line 7, in <module> RuntimeError: Corrupt .mpy file
slender iron
more general question. since this:
<https://learn.adafruit.com/adafruit-airlift-...
It takes space so I'm not sure we want to blanket enable
@idle owl what board are you on?
idle owl
Feather RP2040.
slender iron
k, I can try it once I'm done with email
idle owl
Ok
slender iron
dan will likely see this once he's back from his meeting too
idle owl
Yep.
tulip sleet
This reload thing is awful. This over and over while copying libs to the board. ...
Are you saying 7.2.1 is not working?
i just got out of a non-work meeting
idle owl
Are you saying 7.2.1 is not working?
It's reloading repeatedly on every save and every write. 6 times on saves, copying files, it reloaded every 1/4 second the whole time.
No worries
tulip sleet
it's supposed to stretch out the restart time
that is not the way I want it to work
@idle owl which board are you trying?
idle owl
Feather RP2040
manic glacierBOT
tulip sleet
Feather RP2040
are you using pycharm as the editor, on Mac? I don't see repeated reloads on Linux with emacs or circup
I will try it on the Mac
idle owl
PyCharm.
And iTerm 2, and tio.
Finder did it too though.
So maybe it's a MacOS issue.
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/1oq50yQiocKcPcGAHEyFuRJqNOz5ipXDD62D0aPNctq8/edit?usp=sharing
random junco
I don't know if it helps, but I can reproduce the issue on macOS
tulip sleet
I am trying it with Mu - I see two auto-reloads (not 4)
I am using an M1 Mac Mini
I will instrument the restart loop for myself to see what is going on
idle owl
This is what happens when I edit this code: ```0.489988
0.519988
Code stopped by auto-reload.
soft reboot
Auto-reload is on. Simply save files over USB to run them or enter REPL to disable.
code.py output:
Code stopped by auto-reload.
soft reboot
Auto-reload is on. Simply save files over USB to run them or enter REPL to disable.
code.py output:
Code stopped by auto-reload.
soft reboot
Auto-reload is on. Simply save files over USB to run them or enter REPL to disable.
code.py output:
That last one finally lets me do something.
random junco
I'm on an M1 Macbook, and I copied a font to my feather, and it did it 6 times it looks like, just like Kattni above
idle owl
I'm also on M1.
tulip sleet
the way the code is written, if the writes are <0.75 seconds apart, it should not restart multiple times
so I will verify that
lone axle
@idle owl okay, post meeting things are wrapped up. I am ready any time.
idle owl
@lone axle Ok, BRB, refilling water. Then we can get going.
tulip sleet
I'm on an M1 Macbook, and I copied a font to my feather, and it did it 6 times i...
did you copy with the Finder or in the shell?
random junco
Finder
tulip sleet
I am copying Arial-12.bdf to the drive. On Linux, I get one autorestart. On MacOS, in the shell, I got two
idle owl
Become a user, Dan. Use .... Finder....
tulip sleet
Become a user, Dan. Use .... _Finder_....
Just trying to figure out what is different 🙂
hidden oxide
Become a user, Dan. Use .... _Finder_....
Nooooo
slender iron
it may be the metadata writes that apple/finder does
tulip sleet
it seems to be worse with the RP2040. I tried an ESP32-S2 and a SAMD51 on Mac and I am not seeing multiple restarts
slender iron
maybe writes take longer on rp2040?
tulip sleet
Yes, I may just need to stretch the 0.75 seconds. XIP and writes are getting in the way of each other
maybe
slender iron
it has aggressive interrupt disabling while writing flash that may have delayed supervisor_ticks before
tulip sleet
that would make a lot of sense
We are having dinner guests in a couple of hours and I have to cook. I will do some more work on this later.
slender iron
so you .75 measurement is truer than before
want me to take a look?
aka, increase the reload time on rp2040?
tulip sleet
if you'd like, sure. I did try the RP2040 on Linux and it doesn't seem as bad, so there may be multiple writes: all the junk files macOS writes
slender iron
yup, that's my guess too
I'll take a look shortly
7.2.2 can have this tweak plus re-enabling bleio on matrix portal
tulip sleet
ty!
slender iron
np
idle owl
@lone axle I moved the if START_TIME: indentation out one, and it works now. 🙂 I realised what might be happening, and that fixed it. I'll submit a PR to your repo!
@slender iron Reload using ctrl+D does not reload a bunch, only once. Same out of the REPL using CTRL+D
slender iron
that's good 🙂
I suspect we just need to increase the timeout for reloads on the rp2040
idle owl
Oki
slender iron
(I'm distracted now trying to test my esp change)
idle owl
No worries, simply wanted to let you know I had more data.
slender iron
ty
idle owl
@lone axle Turns out you can't set brightness on a single pixel. So, my code isn't doing what I wanted it to. Debating between leaving it with both keys pulsing, trying to figure out another way to do it, or using the LED Animation library.
I wanted only the time-start key to pulse.
lone axle
Ah, I discovered recently that Dotstars can do individual brightnesses. Hadn't known the differences before.
idle owl
Interesting.
I wonder why.
Maybe I'm doing it wrong. 😄
I was trying to do pixels[1].brightness which failed.
blissful pollen
Ah, I discovered recently that Dotstars can do individual brightnesses. Hadn't k...
Is that a code thing or hardware?
lone axle
For neopixels one option for "lower brightness" would be to use a color closer to black. i.e. 0xff0000 is full bright red, but 0xaa0000 is less bright red. It may get convoluted if you are attempting to puslse it up and down for a fade in / out effect though so maybe isn't the best solution for that.
idle owl
For neopixels one option for "lower brightness" would be to use a color closer t...
Yeah I was thinking the same thing. Easier to do using RGB I think. At least for me.
lone axle
Is that a code thing or hardware?
My understanding is it's a difference in hardware, but I did happen to discover it because I noticed code somewhere that was setting it individually and then further confirmed it in the library code.
ATMakers Bill ended up finding code in the core I think as well that referenced the same behavior. There should be a link and brief discussion somewhere in the #live-broadcast-chat history if you're interested in going deeper through what we found.
blissful pollen
I always used what you just said change the value from 0x00 to 0xFF but will have to take a look at the dotstars thanks
proven garnet
Anybody have a SharpMemoryDisplay? Would love some help on quickly testing a PR for the library!
idle owl
@lone axle Called in a loop, but it doesn't update: pixels[1] = (0, (g + 1) % 255, (b + 1) % 150) What obvious thing am I missing? Is that not how % works? 😄
Now that I think about it, it's not going to be the right color anyway.
it will for part of the time. But never again after that.
If they're updating at the same speed anyway.
I could change it to straight up cyan, which might make this simpler. But I'm still not sure how to do this.
lone axle
Ah, it looks to me like if g and b are incrementing it would work as far as I can tell. Only gotcha I can think of is probably need some time.sleep() or some way to slow it down some, by default it might iterate so fast that it's difficult to see the changes.
idle owl
b/g are 0 before the while loop.
I have time.monotonic() around it.
In the loop. That's how I timed the pulsing before.
It's not iterating though, afaict.
Printing (b, g) is (0, 0).
blissful pollen
Maybe it's my lack of python knowledge but isn't g and b not incrementing? you're just adding one to them
idle owl
Shouldn't they add 1 255 times?
And actually overflow at that point because I haven't added the "if 255, to back to 0" code.
lone axle
And actually overflow at that point because I haven't added the "if 255, to back...
The modulo % would take care of that so it wouldn't be strictly necessary if I understand.
idle owl
Either way, it's not even adding 1. They're set to 0 before the loop, and it prints 0 after that pixel line in my code.
That's what I thought... ok..
lone axle
can you drop it into a gist? I can take a look in context and see if anything jumps out
idle owl
This seems like something I'm trying to do simply, and maybe it's not so simple.
Sure!
I guess it could all be under one if now because indentation changed.
Updated that. Still doesn't work, but at least it's not redundant.
Also going to have to not leave this plugged in or I'm going to be sending timestamps to random things when I get distracted and start pressing toebeans. 😆
FWIW, I tried the LED pulse line outside of the if in only the while True: and it still didn't do anything.
I also understand this would only pulse up. I was going to tackle the other direction once I got this working.
lone axle
https://gist.github.com/kattni/b98b3e0ad159c71100cb3e366cad73f5
Ah I think you'll need to increment b and g inside of this if statement:
if time.monotonic() - PULSE_TIME > 0.05:
pixels[1] = (0, (g + 1) % 255, (b + 1) % 255)
print(b, g)
PULSE_TIME = time.monotonic()
b += 1
g += 1
ya sorry that is what i was (badly) trying to say earlier 🙂
idle owl
Wait, no it doesn't....
😂 It's up to 721 by now.
It was a soft reload that made it look right.
Do I put the % in the loop outside the pixel line?
lone axle
it will probably be blinking from full bright to full dark if I understand correctly. And "fading" up from full dark to full bright.
idle owl
Yep
Oof. Autoreloaded on 253 🤦🏻♀️
Oh!
Code is working, print is incorrect.
Alright.
lone axle
in order to fade back down instead of blink you can use b -= 1 and g -= 1 but you'll also need some logic to tell it when to do which. I would probably make a new boolean fading_up or something and set it to true for increasing and then false once it reaches 255 to start the decrementing. And set it back to true once it reaches 0. At that point you could also get rid of the modulo as well I think.
blissful pollen
you can use like b = (b + 1) % 255 and remove any long term overflow possibility (which would take forever to hit)
idle owl
Ok
idle owl
Nope.
Check the gist above, that's the basic code I'm running. (Modified now, but I'll update it in a minute.)
slender iron
ok, one thing I'm seeing is that the autoreload delay only applies if code.py was interrupted
idle owl
Like by me hitting save? Also when it decides to auto-reload on its own, it reloads a bunch of times. But once every 10 times, it's fine.
slender iron
hrm. I'll definitely get my mac out to try it too
tulip sleet
ok, one thing I'm seeing is that the autoreload delay only applies if code.py wa...
The autoreload flag is checked in the sleep loop, because that's fake sleeping (even though the VM is gone), so that it will autoreload immediately if sleeping. But yes, it should have the delay loop there too, in case there are multiple writes (which there will be). So I bet it's restarting immediately, and then it falls into the actual running check. So need to factor out the delay loop and call it two places, I guess.
slender iron
I'm thinking it might be clearer to have the global just be the last write completion time
tulip sleet
i was thinking of that too, but I was a little worried about the cache
tulip sleet
the cache is flushed explicitly on VM exit, but maybe not on VM entry
slender iron
seems like flushing on vm entry could be good too
tulip sleet
since the ticks are uint_64, it can't be an atomic volatile. so maybe need to make setting it be non-interruptible.
slender iron
or we add some 32 bit tick math
tulip sleet
instead of being a global, it could be a few routines in reload.c that you call to set and check or get it
idle owl
@lone axle I got it working, but it flickers every so often, and when it does, both LEDs flicker, not only the one being pulsed. No consistency on the flickering. When it does, it's a single flicker every so often. Gist updated.
Initially the flickering was because it was hitting -1 or 256, but I fixed that.
Flickering is not consistent though. It's not happening right now. 🤷🏻♀️
The only issue I can foresee with this is if the reloading is this finicky, it's going to lose its elapsed time on reboot. But hopefully that's resolved soon 🙂
I'm not about to figure out how to save the elapsed time in nvm or whatever the right term is. 😄
lone axle
Flickering is not consistent though. It's not happening right now. 🤷🏻♀️
weird. Nothing jumps out to me from the current version as a potential cause of that. Tonight is the first night of an online course I'm teaching for the next few weeks, so I won't have time to get further into it tonight. But I should have some time tomorrow to find and assemble one of the neokey featherwings I have and try it out to see if I can replicate the flicker.
idle owl
weird. Nothing jumps out to me from the current version as a potential cause of ...
No worries! Have fun! I'll keep the gist updated if I make any more changes. Simply let me know when you get around to it.
lone axle
will do
thorny jay
I also made similar suggestion for the learn guide to try to have both ESP32-S2 QT Py more equal even if one came later.
slender iron
@idle owl around still? want to try what I have now?
I haven't tried it on my mac yet though
idle owl
Yes, and very yes!