interesting. it shows everything green but actions to me. I wonder if there was some large scale abuse of actions that started creating wider problems.
#circuitpython-dev
lone axle
dusk mauve
interesting. it shows everything green but actions to me. I wonder if there was ...
gives me a flashback to a cloud cascade failure I had to manage several years ago in a past life. "One repeating user API call ate up all 64 GB of heap? On ALL of the nodes? All 1,200 of them?" π
Making drivers fit SAMD21 memory is super calming by comparison.
lone axle
ah, more non-greens now. Looks like the status page is figuring out how big of a problem there is π
stuck elbow
just when I started coding, github goes down
ember iris
There it goes, they see the outage now
Guess it's time to work on local stuff only? I was hoping to find time to play with some new sensors...
dusk mauve
There it goes, they see the outage now
Whoever's in charge of their incidents finally got the page. I hope their on call crew is well taken care of. Its a tough life. π
ember iris
Hopefully the fix is simple and they can be low staffed through the rest of the weekend
dusk mauve
with a widespread failure like this though a lot of CI/CD systems will start failing and cascade into thousands of build pipelines all over the place.
stuck elbow
if your CI depends on external services, you have bigger problems
dusk mauve
if your CI depends on external services, you have bigger problems
one of my former bosses would disagree, but that's why they're a former boss. π
stuck elbow
one of my former bosses would disagree, but that's why they're a former boss. π
this is why we practice boss fights in games
dusk mauve
I should be eating a late lunch anyway. π Here's their incident page if anyone who also has work to do wants to subscribe/follow along.
warm stump
it's back up but maybe give it a break for a while anyway π
manic glacierBOT
solar whale
yikes -- updating CP -- downloading the broadcom/firmware takes a long time!!
stuck elbow
yup, there is an issue for it already
perhaps a separate script for downloading it would make sense
manic glacierBOT
lone axle
@lunar gull I have submitted a PR that adds a new cell_anchor_point feature to the GridLayout. https://github.com/adafruit/Adafruit_CircuitPython_DisplayIO_Layout/pull/58 If you have a chance it may be good to give this a try (if you need help with figuring out how to try a branch from a PR let me know). This allows you to declare an anchor_point to be used when laying out the content into each cell. using (0.5, 0.5) results in content getting centered within the cells and does not require using that extra layer of Group objects like the example I linked yesterday. The approach from this PR is probably better for long term usage since it allows the user code to be less complex.
I saw you mentioned IconWidget not working as expected. I still need to look into that, I would have expected IconWidget to behave pretty much the same as Label with respect to it's anchor_point and anchored_position, so if it's not then we'll try to figure out why and fix it. If you have a chance it may be worth giving the version from this PR a try with IconWidget to see if it behaves as expected if so it could get you past that issue, although it's still something we'd want to fix in the longer term I think.
lone axle
@gilded cradle whenever you are around next I have a question: Do you know if items in a Palette are meant to be a different type in Blinka_DisplayIO than they are in the core? For example this code:
palette = displayio.Palette(1)
palette[0] = 0xFF0000 # red
print(palette[0])
run on a microcontroller prints this 16711680
But I'm running the same code with Blinka_DisplayIO and seeing it print this: {'transparent': False, 'rgb888': 16711680, 'rgba': (255, 0, 0, 255)}
gilded cradle
<@!478616919721771019> whenever you are around next I have a question: Do you kn...
Yeah, I canβt remember exactly the reason I did that. Perhaps it was to get around a limitation of pillow or something. If you want to try and get it to be the same type and still have it function, that would be cool.
lone axle
Yeah, I canβt remember exactly the reason I did that. Perhaps it was to get arou...
I tested out a change to make it return the "rgba""rgb888"value from the dictionary and it does seem to work in the specific case I tested. I'll test it out a bit more thoroughly and PR it if nothing weird shakes out.
gilded cradle
I tested out a change to make it return the ~~`"rgba"`~~ `"rgb888"`value from t...
Sounds good.
crimson ferry
Does ESP32-S2 use a less-accurate clock source while it's deep sleeping?
(in fake deep sleep, ESP32-S2 drifts by a second over a span of hours; in real deep sleep, it drifts almost a second per minute)
lunar gull
IconWidget positioning
@lone axle I'm going to try that code out in few hours. Stepping away for some TV-time with the wife.
lone axle
Sounds good. Have a good night!
edgy drum
How does one use the mpy-cross tool to bundle a module (as opposed to a single file)?
edgy drum
Thanks so much! It occurs to me now that .mpy is a reference to mircopython. Never made that connection
edgy drum
It is! I've only used CP so any MP references go above my head
lunar gull
I'm quite new to it as well, but I love Python, and I love CircuitPython and everything Adafruit, so here I am. π
edgy drum
we are alike
lunar gull
(I also have some love for SparkFun, because they have some crazy sensors for all kinds of stuff that are all QT)
Sometimes they do not play together, but recently I've seen them collaborate on things, together with Digi-Key, which is another favorite along with Mouser. I sometimes buy a bunch of weird electronic components and even electrical components (I mess around with managed power distribution and power management on my free time for "overlanding").
crimson ferry
more about mpy-cross for CircuitPython here: https://learn.adafruit.com/creating-and-sharing-a-circuitpython-library/creating-a-library
lunar gull
One of the coolest things for Qwiic, though, is the Qwiic to RJ45. It's powered, splits the signal up into 4 so it can be transmitted over greater distances. Let's you make a run of QT wire that's measured in feet instead of inches. It's also perfect because Ethernet cables are generally way sturdier, even at a cheaper level. Sure, it gets a bit pricey when you buy them for a vehicle install because they have to be grease and chemical resistant, but still, waaaaaay better than trying to pull 4 strand wire from a truck cab to the bed.
@crimson ferry does Adafruit have any thing of that sort? I am sorry for using SparkFun prods.
crimson ferry
np, I don't think so... differential for longer distances #help-with-projects message
edgy drum
more about `mpy-cross` for CircuitPython here: <https://learn.adafruit.com/creat...
yeah these instructions convert .py files to .mpy files, I assumed there way a way to convert a module (with its __init__.py and supporting files) into a single .mpy file. that doesnt seem to be how it works though
lunar gull
This one maybe? https://www.adafruit.com/product/4756
lunar gull
yeah these instructions convert .py files to .mpy files, I assumed there way a w...
Yeah, it converts each file.
edgy drum
Using ethernet cables as a "hack" for long range is one of my favorite things
lunar gull
np, I don't think so... differential for longer distances https://discord.com/ch...
Yup, those are the ones. Absolutely amazing. I have all of "sensitive" electronics and the DC-DC charger in the cab of my truck, but there's a big 'ol box in the back corner of my bed that houses switching MOSFETs for all the different DC stuff like lights, fridge, whatever, and I didn't want to have to mount any Arduino/RPi back there, so the QT extenders are clutch.
crimson ferry
we should probably take the non-core-development stuff to another channel
lunar gull
Yes.
manic glacierBOT
main furnace
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
TypeError: 'mapping' argument required
'mapping'?
via Adafruit LED Glasses Driver nRF52840
stuck elbow
@main furnace where did you get that library?
main furnace
It's in the CP build I got from S3: Adafruit CircuitPython 7.1.0-beta.0-153-g27e8a16bf on 2021-11-26
stuck elbow
I don't see is31fl3741 in the frozen directory
and looking at the mpconfigboard.mk for the glasses, they don't freeze any libraries on it
looking at https://circuitpython.readthedocs.io/projects/is31fl3741/en/latest/api.html there is no "mapping", but neither is there "scale" or "gamma"
main furnace
According to https://github.com/adafruit/circuitpython/pull/5584 it is only compiled for the LED Glasses controller. Don't know what that means.
stuck elbow
ah, it's the framebuffer one
main furnace
Yeah. Reading further in the pull request I found this:
from map import glassesmatrix_ledmap
is31 = is31fl3741.IS31FL3741(width=54, height=15, i2c=i2c, scale=True, gamma=True, mapping=glassesmatrix_ledmap)
Guess I need to find map.py then.
stuck elbow
main furnace
Guess that's what I get for playing on the bleeding edge. π
blissful pollen
No problem if you have any questions let me know, I'm still refining the library (I want the eye rim lights to work too)
stuck elbow
@blissful pollen why list and not a tuple, or even int array?
blissful pollen
<@!228322543507865601> why list and not a tuple, or even int array?
Just what I chose and it was copied from Arduino and my lack of experience in pure python
I looked at bytearray but for int16 values it was causing my issues with big/little endian
daring pumice
Good morning everyone π
@tulip sleet, wondering if you have a few minutes at some point today to discuss the nrf52833 issue - I've been going over some the adjustments made by xobs, and wanted to make sure I'm thinking understanding them correctly.
No rush - just getting rolling.
thinking/understanding*
blissful pollen
<@!228322543507865601> why list and not a tuple, or even int array?
Offhand is the tuple (or int array) more efficient? Probably isn't a hard change if its worthwhile and can stick it on a future todo list
stuck elbow
yeah, tuple is not modifiable, so saves a bit on that, an array would be the most closely packed data structure for this, but you need to import array for it
main furnace
Whoa, it's trying to display the console on the LED Glasses!
stuck elbow
I use 16-bit arrays for palettes, for instance
tulip sleet
Good morning everyone π
<@!329766224093249548>, wondering if you have a few mi...
Go ahead and post your thoughts here or in the issue.
daring pumice
Go ahead and post your thoughts here or in the issue.
Fair enough - is it alright if I get back to you a little later? Setting up our sprint meeting for tomorrow and answering emails atm.
blissful pollen
Whoa, it's trying to display the console on the LED Glasses!
Yes it does (badly but you see Blinka ha)
blissful pollen
yeah, tuple is not modifiable, so saves a bit on that, an array would be the mos...
Ah thanks I'll take a look at that.
stuck elbow
I know this is penny-pinching, but saves 2 bytes per entry
I'm still living with the samd21 boards
blissful pollen
It is copied internally (as it has to be long living for displayio). May be worth putting that in the help that you can deinit the map after importing it.
stuck elbow
oh, in that case don't bother
blissful pollen
I get you with the samd21 stuff. I know once you give me more memory I tend to forget that stuff
stuck elbow
it will just disappear once no variable refers to it
hm, but if it's in a an imported module, it will live forever by default
and in the long-lived memory too
tulip sleet
Fair enough - is it alright if I get back to you a little later? Setting up our ...
np, at your pace
daring pumice
np, at your pace
Thanks much π
turbid radish
?serverinfo
digital shoreBOT
adafruit
daring pumice
@tulip sleet
For the nrf52833:
I think the issue is that SD_FLASH_SIZE has two different definitions depending on SoftDevice version:
v6.1.0: 0x25000
v7.1.0: 0x26000
If the make is done with v7.1.0, it makes the start of the application space 0x27000 - that's 4KB past where it should be. This accounts for the error I previously reported on GitHub.
Thoughts?
Bear in mind that I'm getting the SD address start from Nordic's docs: https://infocenter.nordicsemi.com/index.jsp?topic=%2Fsdk_nrf5_v16.0.0%2Flib_bootloader.html
Scroll down to the Memory Layout Section.
tulip sleet
<@!329766224093249548>
For the nrf52833:
I think the issue is that SD_FLASH_SIZ...
The SD is larger, so the application space should be larger by 4KB, 0x1000. But are you saying the application start address assumes 0x25000. In the bootloader or in CircuitPython?
daring pumice
The SD is larger, so the application space should be larger by 4KB, 0x1000. But ...
No, I'm saying that the declared SD size is bigger than Nordic specifies is allowable. According to their docs, they expect the firmware to start at 0x26000. However, if the SD size is 0x26000, then the firmware start is pushed to 0x27000 - that's too far.
tulip sleet
if you look in ports/nrf/build-pca10100/common.ld, it looks like the appropriate values have been substituted in the templated LD files
tulip sleet
No, I'm saying that the declared SD size is bigger than Nordic specifies is allo...
I think the bootloader doc you are pointing to is assuming the old SD size. They increased the SD size, that's where we got the 0x26000 from. But something else may be assumingthe wrong size
or the wrong start address
daring pumice
I hear you. Not sure what to make of it - not strong enough here.
Can I send you a Screenshot here?
tulip sleet
sure - I am about to be in a meeting at the top of the hour
daring pumice
kk 1 sec
tulip sleet
I have been aware of this size increase for years, and added SD_FLASH_SIZE to handle it. But there may be a bug somewhere.
tulip sleet
I did not think the SD size varied by processor. Maybe I made a mistake. I may have been going by the '840 size.
try reducing SD_FLASH_SIZE temporarily and see if it works
daring pumice
I did not think the SD size varied by processor. Maybe I made a mistake. I may h...
I hear you, and seem to remember this being an issue on the nordic Dev boards. I'll find the site.
daring pumice
try reducing SD_FLASH_SIZE temporarily and see if it works
I have, but this creates knock-on issues -> there's an overlap when merging.
Ah - seems like a documentation error: https://devzone.nordicsemi.com/f/nordic-q-a/56730/s140-v7-0-x-application-area-address-with-nrf52833
You were right, should start at 27000
tulip sleet
we were thinking of going to SD 7 for '840, but it caused a follow-on effect, needing to update the bootloader/SD. We could include the SD in the CPy build, but haven't done that yet. The bootloader used to need the SD to function, but I think that requirement has been removed (or maybe we just had that as a goal, but it isn't true yet).
tidal kiln
@idle owl hey, do you know anything about if this lib:
https://github.com/adafruit/Python-Thermal-Printer
has been replaced by this?
https://github.com/adafruit/Adafruit_CircuitPython_Thermal_Printer
tulip sleet
@onyx hinge ok, it was something else I was assigned, not ESP32SPI issues
onyx hinge
<@!426200547024961539> ok, it was something else I was assigned, not ESP32SPI is...
@slender iron let me know if you recall which bug this was
idle owl
<@!330227457296957440> hey, do you know anything about if this lib:
<https://git...
No idea.
tidal kiln
np. i'll ask. i'm not sure either.
idle owl
np. i'll ask. i'm not sure either.
Sounds like a good plan.
solar whale
What version of "aarch64-none-elf-gcc" is needed for the broadcom builds
lone axle
@lunar gull I added an item to the "in the weeds" section of the meeting notes to discuss the board module in the stubs. If you're around, we're happy to have you join in or you can follow up with the recording after.
idle owl
@slender iron Remember, put yourself at the top of the list and go through from the beginning.
solar whale
thanks! -- sorry -- did not meant to interrupt.
onyx hinge
@idle owl did we change how we're doing that? is it host, then top of alphabet instead of host, then following in alphabet?
idle owl
Yes.
onyx hinge
π
idle owl
Updating the instructional doc now.
onyx hinge
some other jeff π
gilded cradle
It's 
CircuitPython Monday at Adafruit! Use the code ADATY to get 20% off CircuitPython and 15% off items storewide!
idle owl
I keep forgetting to mention this. Folks: when you click around in the notes document, it displays a little "flag" on your cursor which can block a set of characters to the right of your cursor. Please try to avoid clicking around in an active section while someone is trying to read it because it can block part of what they're trying to read off. Thanks!
thorny jay
I am just listening (no notes)
onyx hinge
@stuck elbow which board(s) are you trying camera on? the adafruit grand central I have basically abandoned, and I concentrated on rp2040 and esp32-s2. all my recent work has only been tested on esp32-s2 only. the samd51 parallel capture peripheral was haunted for me.
stuck elbow
@onyx hinge esp32-s2, but with my own breakout, which has issues
onyx hinge
okay, good luck when you get your new modules!
stuck elbow
@onyx hinge I have a ready breakout, but it misses either mclk or pclk pin, because it only has one clock pin
so I can't use it
onyx hinge
drat
solar whale
You're welcome!
onyx hinge
[brb, my computer audio freaked out]
idle owl
"How do I get into programming?" - https://kattni.com/how-do-i-get-into-programming
onyx hinge
[back]
idle owl
@blissful pollen I should really put you in touch with my partner who a. wrote pixelbuf, and b. was working on getting the LED Animation library working on the IS31 breakout and glasses, and has at least brainstormed getting all of that going. If you're interested.
blissful pollen
<@!228322543507865601> I should really put you in touch with my partner who a. w...
Sure thing. My plan was to start writing some stuff to see how it worked and what others thought
idle owl
Sure thing. My plan was to start writing some stuff to see how it worked and wha...
Keen, tag me when you do, and I'll tag her in.
lone axle
@blissful pollen if you do every find a way to have subclassed native classes I am interested in this as well. I think there are some nifty things that could be done with subclasses of displayio.Group. But I haven't looked into doing it within the core at all yet.
blissful pollen
<@!228322543507865601> if you do every find a way to have subclassed native clas...
If you want to talk about it later (even after in the weeds) let me know. I learned quite a lot about how classes are handled and potential ways to expand it
onyx hinge
@solar whale yes there's a particular version of mkfs.fat that is needed
lone axle
@blissful pollen I'll get in touch when I have a bit more time. We can make a Discord thread for it maybe so other folks can find it in the future more easily.
blissful pollen
<@!228322543507865601> I'll get in touch when I have a bit more time. We can ma...
Sounds good just let me know. Itβs all in my head for now at least
onyx hinge
@solar whale from the earlier discussion I think it turns out the 'secret' is only in the github workflows file ```yml
- name: Install mkfs.fat
run: |
wget https://github.com/dosfstools/dosfstools/releases/download/v4.2/dosfstools-4.2.tar.gz
tar -xaf dosfstools-4.2.tar.gz
cd dosfstools-4.2
./configure
make -j 2
cd src
echo >>$GITHUB_PATH $(pwd)
New Blinka art dropped! Happy little circuits!
solar whale
<@!343570430042308609> from the earlier discussion I think it turns out the 'sec...
thanks !
idle owl
I'm so excited about these. There's a bunch!
onyx hinge
"availibility: <platform name>"
if stubs or py files can have conditionals then ... ```python
if os.environ['CP_BOARD'] == "example_board_1":
D2: microcontroller.Pin
happy little blue smoke
onyx hinge
.. that doesn't work .. this is an error to mypy: ```python
import os
class Pin: pass
if os.environ.get('CP_BOARD') == 'example_board_1':
D2: Pin
D3: Pin
if os.environ.get('CP_BOARD') == 'example_board_2':
D2: Pin
D5: Pin
`pinz.pyi:8: error: Name "D2" already defined on line 5`Hah, even Blinka Ross can't avoid visits from Sparky.
stuck elbow
every machine can be a smoke machine
lone axle
.. that doesn't work .. this is an error to `mypy`: ```python
import os
class Pi...
the code would be a bit more complex. But if it does support conditionals there could be a single larger if statement for each pin that includes checks for each device that does contain that pin. Then each pin would get defined only once (or zero times if they are set to a device without it).
onyx hinge
[4.2 is in debian bullseye so it should be in the NEXT ubuntu]
import os
class Pin: pass
if os.environ.get('CP_BOARD') in ('example_board_1', 'example_board_2'):
D2: Pin
if os.environ.get('CP_BOARD') in ('example_board_1',):
D3: Pin
if os.environ.get('CP_BOARD') in ('example_board_2',):
D5: Pin
``` this is accepted by mypy but D2/D3/D5 are 'available' for type-checking regardless of the CP_BOARD environment variablethat feeling when you are googling for the information, and all you find are your own posts
idle owl
ADATY
onyx hinge
https://raspberrypi.stackexchange.com/a/37167 are there two L2 caches? π€―
idle owl
lunar gull
<@!438791522285256715> I added an item to the "in the weeds" section of the meet...
Apologies, I missed the meeting today.
idle owl
No apology needed π
lone axle
Apologies, I missed the meeting today.
Yep, no worries. There is a recording that you can check out if you want. Some great ideas came from the discussion near the end.
lunar gull
Yep, no worries. There is a recording that you can check out if you want. Some g...
Yay! I'll check it out.
solar whale
woohoo! ```rw-rw-r-- 1 jerryneedell jerryneedell 13355997 Nov 29 15:20 firmware.disk.img.zip
-rwxrwxr-x 1 jerryneedell jerryneedell 1576712 Nov 29 15:20 firmware.kernel8.img
but does it sit in the box?
lone axle
Another idea that just came to mind: make commands can include arguments I presume. i.e. make BOARD=circuit_playground_express Maybe there could be an optional argument added to the make stubs so someone could do something like make stubs BOARD=circuit_playground_express and this would generate the stubs with board being specific to that device. The generic one published to pypi could still contain everything (with a warning that some pins may not be available on every device). But if folks want to go further they could build their own stubs from the core repo and then install them.
slender iron
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/1OH5tU-lCvDNS0vXRsEfyKjiey2V1zjqGjFSHcGmFSH4/edit?usp=sharing
slender iron
Here is the notes document for Mondayβs CircuitPython Weekly meeting. It is at t...
lone axle
I didn't figure out if possible / how to use conditionals. But did get a warning note showing up in the ID for the usage of a given pin. I'll tinker it wit some more this week.
solar whale
@slender iron The deep dive notes are great -- sorry it took me so long to start using them!
slender iron
np, did you find the repo with them?
solar whale
yes
solar whale
it was super to just click on it and have the video pop up!
slender iron
hugs to @idle wharf
idle owl
I think my notes were in the doc tho
They were and Scott read them off. Thanks for putting them in!
slender iron
@onyx hinge this is the other issue I'm thinking of: https://github.com/adafruit/circuitpython/issues/5462
idle owl
Ok, so wait..... I forget this every time. The readonly in storage.remount() is referring to the status of the board flash to CircuitPython, right?
So True means the computer can write to it, and False means CircuitPython can write to it. Right?
lone axle
Ok, so wait..... I forget this every time. The `readonly` in `storage.remount()`...
It's documented in this guide page now: https://learn.adafruit.com/circuitpython-essentials/circuitpython-storage there isn't a header near it so you'll have to scroll down a little, it's under the first code example:
lone axle
You're welcome. I always end up referring to it as well so I know right where to find it. π
idle owl
That's why I documented it. Because I never remember. And then I forgot I documented it, apparently.
blissful pollen
There really was no plumbing / intention to let native classes subclass each other hmm π
idle wharf
it was super to just click on it and have the video pop up!
Woot! So glad you found that useful. I've used it several times myself.
idle owl
@tulip sleet OK, so in pins.c.... Whichever pin comes first is the "preferred" pin, right? My pin script is returning non-silk pin numbers, and I think it's because of the order in pins.c, which I would prefer to update if I'm remembering all of this correctly.
tulip sleet
<@!329766224093249548> OK, so in pins.c.... Whichever pin comes first is the "pr...
I don't think we are consistent about the "preferred pin" being first.
idle owl
Is it alright to fix the Feather ESP32-S2 then?
tulip sleet
sure
slender iron
printing the pin object will use the first name
idle owl
OK, yeah, that's what I meant.
Couldn't think of the terminology.
Should be silk pin.
tulip sleet
there might need to be a big pass through all the pins.c to vet them, though if they are mostly right, maybe not
idle owl
Not sure how to automate that, but I'm not great with automating things like that.
Ooof. @slender iron Should the */broadcom/firmware be taking ages to clone?
Making sure it's not a GitHub- or me-issue.
There it goes.
Must simply take a while.
Wait.
It failed.
Oi, my disk image isn't big enough it seems.
Guess it's time for a new one.
Wait, seriously? How big is the broadcom stuff?
idle owl
I mean presumably I'll eventually need it. How big is your disk image for CircuitPython?
--depth 1 worked though.
idle owl
Oh fair enough. I guess I don't know what depth is doing.
tulip sleet
@slender iron draft for 7.1.0-beta.1 is in. I added the broadcom port to the alpha list, and pointed it out in "Notable additions".
See if you want to add anything.
I don't think there's anything immediately on the agenda to add. There is a small translations PR, but we can wait on that.
slender iron
depth is the number of commits to fetch on a branch starting with the latest and going back in time
idle owl
Ah. Thanks.
slender iron
k, will take a peek @tulip sleet
usually the history is small enough that getting it all is fine
onyx hinge
@slender iron OK it would take me time to get up to speed on ble hci. it's using uart, not spi, right? still want me to look into it (tomorrow)?
slender iron
@tulip sleet will have a better idea than me
tulip sleet
i can take a look instead. This looks like a simpler connection problem of some kind.
it sets it all up using a library I wrote, etc.
tulip sleet
(adafruit_airlift)
idle owl
See if you want to add anything.
Can add this to the release if it's not troublesome. https://github.com/adafruit/circuitpython/pull/5626
tulip sleet
Can add this to the release if it's not troublesome. https://github.com/adafruit...
Will def add
slender iron
@tulip sleet release notes look good to me. you'll need to add the rpi boards to the new boards list too
slender iron
ya, I'll add notes to the board page
idle owl
@tulip sleet Huzzah! ```Auto-reload is on. Simply save files over USB to run them or enter REPL to disable.
code.py output:
Touch on: A4
Touch on: A5
Touch on: D10
Touch on: D11
Touch on: D12
Touch on: LED
Touch on: SDA
Touch on: SCL
Touch on: D5
Touch on: D6
Touch on: D7
Touch on: D9
LED should be D13, but, I'd rather have LED be the primary pin, so... too bad.
tulip sleet
I will start a release as soon as the merges finish.
idle owl
I will start a release as soon as the merges finish.
Right on. Thanks so much!
Probably should have swapped D7 for I2C_POWER_INVERTED, but the guide refers to pin 7 on the pinouts page, so..... I was leaving the power-related pins alone, and glossed over that.
daring pumice
Re: NRF52833
Having gone over the memory mappings, I'm still left with the only clue I have as to why CircuitPython won't deploy onto the 833.
nrfjprog -f NRF52 --program firmware.hex -erasechip --verify --log Parsing image file. Verifying programming. ERROR: [ nRF52] - Data does not match in address range [0x00027000-0x00027100] (Flash) ERROR: [ nRF52] - Expected byte value 0x02 but read 0x00 at address 0x00027002. ERROR: [ nRF52] - Flash verification failed. ERROR: [ nRF52] - Failed while verifying file firmware.hex. ERROR: Write verify failed.
If you recall, the reason why I suspected that the SD had gone into the application space is because of this error... now I'm wondering if it's the reverse: The flasher is reading 0x00, but expects 0x02.
Curious how I can find out what it's expecting (i.e. what's supposed to be in that address) vs what it's getting (what's actually in that address).
Anyone have any thoughts where I can look, here?
Note: I've tried changing -erasechip to -erasesector - no difference.
tulip sleet
The .hex file is human readable
idle owl
Re: NRF52833
Having gone over the memory mappings, I'm still left with the only...
(Tangentially related, three backticks on either end of a code block, makes a codeblock - see the #welcome channel for details) codeblock
daring pumice
Ok. I have it open in vscode. Am I correct in assuming that this file should be loaded to the 833 starting at 0x27000?
daring pumice
(Tangentially related, three backticks on either end of a code block, makes a co...
Nice π Thanks so much π
tulip sleet
Ok. I have it open in vscode. Am I correct in assuming that this file should be ...
yes, it should
slender iron
@onyx hinge what ulab branch should cp be tracking?
onyx hinge
@slender iron main or master if they haven't shifted yet, I think. I haven't kept up with news from 'over there'.
slender iron
kk, cp is still set to use the adafruit fork
slender iron
will fix it in my upcoming submodule pr
onyx hinge
so it is...!
as far as I can get git to tell me, we don't have any commits that they don't have in master.
onyx hinge
we had to take one compatibiltiy change, maybe for mp 1.17 merge, which must have been taken upstream
slender iron
ya, it is in upstream
onyx hinge
but they haven't released since 3.3.1 so there's not really anything to update to
slender iron
ya, I'm not changing it. just fixing the upstream url
idle owl
I tried to build for ESP and .... I did not succeed. It's fine, I grabbed the build off of the PR, but I would have liked to have done it myself. Is it super bonkers and I should avoid it? Or am I simply missing something simple? (I installed cmake and ninja after running into those errors...)
tulip sleet
@slender iron I don't need this submodule fix for 7.1.0-beta.1, I assume?
idle owl
@slender iron The make fetch-submodules concept seems like a pretty significant change. We will need to document it somewhere obvious, I would think.
slender iron
nope. it's just a tooling thing
tulip sleet
i have to put it in "Building CIrcuitPython"
slender iron
@idle owl I figured we'd update the build guide once it is merged in
idle owl
<@!330227457296957440> I figured we'd update the build guide once it is merged i...
Ok, sounds good.
slender iron
you can do the old way but it'll fail on the rpi firmware folder
it won't hurt anything
idle owl
Ah, ok. But folks will run into the error. And we will help. Heh.
tulip sleet
once kattni's pr merge is finished, I will tag the release. @slender iron I can do the branching stuff later
then it has to build (twice) before I do the actual announcements
slender iron
@tulip sleet we should include https://github.com/adafruit/circuitpython/issues/5628 in known issues in the notes
tulip sleet
<@!329766224093249548> we should include <https://github.com/adafruit/circuitpyt...
I will add it. I have the blog post and forum post set to go for later.
ornate breach
I just saw some of the new Blinka
Bob Ross blinka might be my favorite
Also Blinka teaching circuits is also adorable
tulip sleet
@slender iron tools/ci_fetch_deps. py is breaking when attempting to do the circuitpython-org update
this is a new file. It's trying to do the "website" port, which it's not prepared for
the placement of the error is a bit misleading in the logs
python tools/ci_fetch_deps.py website refs/tags/7.1.0-beta.1
target ends up being "website":
else:
p = pathlib.Path(".").glob(f"ports/*/boards/{target}/mpconfigboard.mk")
if not p:
raise RuntimeError(f"Unsupported target: {target}")
config = list(p)[0]
should probably be p = list(...) otherwise the if not p is testing a generator
but that's a secondary bug
I'm not sure why it's doing this at all for the website CI action
I think I can run this by hand. I did that once before
- name: Get CP deps
run: python tools/ci_fetch_deps.py website ${{ github.ref }}
in create_website_pr.yml should probably not be there
slender iron
sorry about that @tulip sleet. I think I added it because of a previous submodule thing
tulip sleet
np, I was able to run the circuitpython-org update as myself
i just had to generate a temp token.
orchid basinBOT
manic glacierBOT
tulip sleet
all CPy builders: note there is now a new way to fetch submodules, which prevents fetching all the commits in some large submodules (in particular the RPi broadcom firmware repo, which is very large). You do cd circuitpython; make fetch-submodules. See https://learn.adafruit.com/building-circuitpython/build-circuitpython#fetch-the-code-to-build-2986719-1.
lunar gull
Can someone help explain why the CircuitPython alarm module is not available for the SAMD5x boards like the Feather M4 Express? That chip clearly has support for several different sleep modes as well as ways of also maintaining RTC states in certain sleep modes. Is that something that can be added, and where do I start digging to come up with a PR, perhaps?
edgy drum
It actually sounds like alarm was added for SAMD chips in the beta that was released today: https://github.com/adafruit/circuitpython/releases/tag/7.1.0-beta.1
lunar gull
aww yeah
lunar gull
It actually sounds like `alarm` was added for SAMD chips in the beta that was re...
Thanks, Dan!
Alright, will mess around with that tomorrow then, will have to upgrade both the RP2040 and the M4 to 7.1 b1.
lunar gull
I am trying this now with the latest firmware on the M4, but I am getting ImportError: No module named 'alarm'. I'll follow up in the morning. But, I have to go to bed now.
tulip sleet
I am trying this now with the latest firmware on the M4, but I am getting `Impor...
alarm was mostly implemented for SAMD51, but was only turned on for a few SAMD51 boards (some satellite boards). To turn it on for all boards requires more work; I am looking at that now.
idle owl
Bob Ross blinka might be my favorite
Happy little circuits! (I'm really happy about this one too.)
idle owl
**all CPy builders:** note there is now a new way to fetch submodules, which pre...
For my own clarity... and yes I looked at the guide but I want to make sure I get it.... make fetch-submodules replaces the entire git submodule sync --quiet --recursive && git submodule update --init command?
ornate breach
Iβm glad my Blinka teaching idea made it through the gauntlet π
I like the way it turned out, all of them really
idle owl
Iβm glad my Blinka teaching idea made it through the gauntlet π
Actually, that one was me. I wanted her drawing something super complicated on a chalkboard. That's what we got. That's only half the Blinkas, there is another set coming soon.
I think your backpack one made it in though?
If that was you..
ornate breach
Yeah, back pack
idle owl
Yeah I believe I kept that one in the list.
Should be part of the second set if they went with it. Which I'm pretty sure, based on this, that my list is canonical.
ornate breach
I had mentioned in the back pack explanation about her impact on education so I thought there was a correlation π
idle owl
I had mentioned in the back pack explanation about her impact on education so I ...
Ah right! Nope, different things entirely.
Her writing in a notebook wearing half-moon reading glasses went with my complicated things on a chalkboard idea.
ornate breach
Well, I love the blinka Bob Ross and drawing a circuit
idle owl
Basically, I wanted her being a smartypants.
ornate breach
They are super adorable
ornate breach
Absolutely π
idle owl
Bruce pretty much always nails it. I might have had feedback on something once or twice, and it was preference stuff, not "fix this pls" type feedback.
ornate breach
Some people just absolutely nail their craft
Itβs always exciting to see what those people do and how they do it
idle owl
For sure
half radish
GM. <noob>Is there an official lookup to map specific feathers to the available atmel-samd/boards for ada boards? (I'm compiling for a bluefruit sense, and want to be sure I get asyncio)</noob>
idle owl
@half radish I don't quite understand the question. You want to see what's available for the Bluefruit Sense, and add asyncio if it's not available?
Hmm. @tulip sleet Why isn't asyncio or some related naming scheme showing up in the support matrix page?
I'm looking at "latest".
Is it not actually part of CP?
tulip sleet
asyncio is a Python library. It really only requires async/await support, and they are keywords
tulip sleet
GM. <noob>Is there an official lookup to map specific feathers to the available ...
async/await is turned on for all boards except SAMD21 and similar small-flash boards
idle owl
Thanks Dan
idle owl
@tulip sleet I just misread it as "Competitive Multitasking" and now I need that to be a thing.
tulip sleet
that's a sport where you watch TV, cook, talk on the phone, and solve a crossword puzzle at the same time. First one to finish the puzzle with fewest mistakes, plus placing first the British Baking Show simultaneously, wins
idle owl
that's a sport where you watch TV, cook, talk on the phone, and solve a crosswor...
π Well played.
orchid basinBOT
manic glacierBOT
main furnace
rpi-zero2w: vectorio shapes are always black. The pixel_shader palette is ignored. Not yet implemented, or should I write an issue?
lunar gull
@tulip sleet you magnificent person, you! That was quick, and also, I think I love you.
tulip sleet
well, it's not done yet, and it's my job π
lunar gull
A lot of people do jobs. But very few immediately address an issue that was reported last night at like 1am in the morning. π
lunar gull
Looks like there's only one board that's failed so far in the builds for the PR.
(but there's still 127 queued, lol)
half radish
async/await is turned on for all boards except SAMD21 and similar small-flash bo...
Interesting... I'm getting mportError: no module named 'asyncio' after loading my adafruit-circuitpython-feather_bluefruit_sense-en_US-7.1.0-beta.1.uf2
tulip sleet
Interesting... I'm getting `mportError: no module named 'asyncio'` after loading...
asyncio is a CircuitPython library. You have to get it from the library bundle or use circup to put it on the board.
half radish
`asyncio` is a CircuitPython library. You have to get it from the library bundle...
What is the directory/file in lib/ for the 7.x bundle? I pushed everything from adafruit-circuitpython-bundle-7.x-mpy-20211127/lib and still have the issue.
tulip sleet
ack, it's supposed to be in the bundle. I added it. In the meantime, download from here: https://github.com/adafruit/Adafruit_CircuitPython_asyncio/releases
don't push everything, that's too big
half radish
ack, it's supposed to be in the bundle. I added it. In the meantime, download fr...
ACK
daring pumice
Wondering if anyone has successfully deployed to an nrf52840?
Also, morning everyone π
idle owl
Wondering if anyone has successfully deployed to an nrf52840?
Deployed what?
daring pumice
circuitpython π
idle owl
All of us? I'm confused, I think, about what you mean by "deployed".
idle owl
Or I'm not confused by it, and I'm confused about the question in general.
daring pumice
Put another way, is anyone actively running circuitpython on an nrf52840?
idle owl
Put another way, is anyone actively running circuitpython on an nrf52840?
Absolutely. What's your issue?
daring pumice
No issue - considering using this board over the 833 since the 840 appears allow circuitpython to be easily flashed onto the device π
idle owl
Ah yeah, it's as easy as it is with any other CP-compatible board.
daring pumice
I appreciate it, @idle owl π
tulip sleet
Adafruit doesn't sell any '833 boards, so they don't get as much constant testing as our own '840 boards. Hence the issues people have been seeing with the '833 builds.
daring pumice
@tulip sleet I completely get it, and this isn't a complaint - sorry if it came off that way.
half radish
ack, it's supposed to be in the bundle. I added it. In the meantime, download fr...
@tulip sleet FYSA: dafruit-circuitpython-asyncio-7.x-mpy-0.5.0 β― ls -lai total 12 1447675 drwxrwxr-x 3 tiny tiny 4096 Nov 30 08:46 . 1325592 drwxrwxr-x 5 tiny tiny 4096 Nov 30 08:46 .. 1447688 drwxrwxr-x 2 tiny tiny 4096 Nov 30 08:46 examples <- only examples are present... no mpy. I checked the other bundles from the release and it's all just examples.
daring pumice
You guys have been immensely kind and patient with me, but my project is losing time on this. At this point I'm going to suggest to my team that we move to a board that's well supported if we want to proceed with CircuitPython π
But I can't thank you enough π
tulip sleet
<@329766224093249548> FYSA: `dafruit-circuitpython-asyncio-7.x-mpy-0.5.0
β― ls -l...
yes, it's confusing, it was added to the bundle, but it wasn't in the zip. I'm trying to track that down; very odd
tulip sleet
You guys have been immensely kind and patient with me, but my project is losing ...
np, the '833 builds were contributed by other folks, so we don't necessarily try them ourselves before merging.
half radish
yes, it's confusing, it was added to the bundle, but it wasn't in the zip. I'm t...
Do you want me to file an issue for you?
tulip sleet
no need, i'm already working on it
manic glacierBOT
solar whale
can anyone else who has built the broadcom port confirm taht this looks normal jerryneedell@jerryneedell-ubuntu-macmini:~/projects/circuitpython/ports/broadcom$ make BOARD=raspberrypi_pi4b Use make V=1, make V=2 or set BUILD_VERBOSE similarly in your environment to increase build verbosity. mkdir -p build-raspberrypi_pi4b/genhdr GEN build-raspberrypi_pi4b/genhdr/moduledefs.h QSTR updated aarch64-none-elf-objcopy -O binary build-raspberrypi_pi4b/kernel8.elf build-raspberrypi_pi4b/kernel8.img cp build-raspberrypi_pi4b/kernel8.img build-raspberrypi_pi4b/firmware.kernel8.img 0+0 records in 0+0 records out 0 bytes copied, 0.000169497 s, 0.0 kB/s mkfs.fat 4.2 (2021-01-31) adding: build-raspberrypi_pi4b/circuitpython-disk.img (deflated 95%) in particular , the 0 bytes copied
the files are all there ```-rw-rw-r-- 1 jerryneedell jerryneedell 13355984 Nov 30 12:07 firmware.disk.img.zip
-rwxrwxr-x 1 jerryneedell jerryneedell 1576712 Nov 30 12:07 firmware.kernel8.img
-rw-rw-r-- 1 jerryneedell jerryneedell 16327 Nov 30 12:07 firmware.objs
drwxr-xr-x 3 jerryneedell jerryneedell 4096 Nov 30 12:06 genhdr
-rwxrwxr-x 1 jerryneedell jerryneedell 7625080 Nov 30 12:07 kernel8.elf
-rw-rw-r-- 1 jerryneedell jerryneedell 4589398 Nov 30 12:07 kernel8.elf.map
-rwxrwxr-x 1 jerryneedell jerryneedell 1576712 Nov 30 12:07 kernel8.img
OK, cool. Thanks
solar whale
cool ! ```
soft reboot
Auto-reload is on. Simply save files over USB to run them or enter REPL to disable.
code.py output:
Code done running.
Press any key to enter the REPL. Use CTRL-D to reload.
Adafruit CircuitPython 7.1.0-beta.1-4-g8d6957b79 on 2021-11-30; Raspberry Pi Zero 2W with rp3a0
Nicely done!
solar whale
@slender iron when you have a chance, can you post a link to your "debug hat" on oshpark.
solar whale
Nicely done!
It's all @slender iron
idle owl
I know he did the work for it, but it sounded like getting it going wasn't entirely straightforward. So you get some kudos for getting through it. π
solar whale
Thanks! It took a few tries
slender iron
@solar whale https://oshpark.com/shared_projects/fBq76nP9
solar whale
<@!343570430042308609> https://oshpark.com/shared_projects/fBq76nP9
Thanks! May come in handy... plus your boards are fun to build!
lone axle
@daring pumice there was recently an enhancement made to Circuitpython.org that allows you to filter downloads by processor family. If you click that filter button on the Downloads page and then scroll down a bit there is a section with processor family.
Once you select the one(s) you want you can close the filter again with the X in the top right and you'll be left with just devices that use that processor:
very helpful if your interested in seeing what devices are out there today for nrf52840 or another given processor.
tulip sleet
just when you're about to give up!
daring pumice
lol yup
tulip sleet
@onyx hinge is this related to what you are debugging right now? https://forums.adafruit.com/viewtopic.php?f=60&t=185718
onyx hinge
<@!426200547024961539> is this related to what you are debugging right now? <htt...
No, that wasn't on my radar.
solar whale
@slender iron do you have a few minutes to voice chat about the broadcom port usage?
slender iron
sure!
solar whale
@tulip sleet Do you know offhand if "time.monotonic()" will "reset" after a "deep sleep" -- that is, with the resolution of the time.monotonic() increment get reset by a deep sleep or will it keep accruing?
tulip sleet
I think it may vary by board. @slender iron has talked about always resetting it for consistency.
solar whale
OK -- it's relevant to a question that just came up https://github.com/adafruit/Adafruit_CircuitPython_RFM9x/issues/67
I'll do some testing...
timber mango
@solar whale At least it does not seem to reset on the platforms i've tested it on, even when pressing the reset button.. i assume that would be kinda the same?
timber mango
(that would be nRF)
idle owl
@tulip sleet Should we fix the bundle instead of fighting with getting your temp fix passing?
Or is it a simple fix...
To make it pass.
orchid basinBOT
tulip sleet
<@!329766224093249548> Should we fix the bundle instead of fighting with getting...
I'm not sure it's fixable to make the 6.x mpy-cross work on that source code. We are still building 6.x bundles until 8.x, is that right?
idle owl
I'm not sure it's fixable to make the 6.x mpy-cross work on that source code. We...
Or until 7.1 maybe.
But yes.
We're still building them for now.
tulip sleet
The mpy-cross problem will occur with either a comprehensive or a quick fix to include asyncio
tulip sleet
I may be stuck. I don't think I can conditionalize it, because it's kind of a syntax error, not just some new function that's in 7.x
btw, what do you think of FoamyGuy's suggestion about renaming it adafruit_asyncio? I didn't do that because it was so fundamental
but we have adafruit_hashlib and adafruit_requests, etc.
idle owl
Question about that.
Is it meant to be a literal drop-in for the Python version?
With literally no changes to the code
or does it not work that way.
tulip sleet
it is a subset, in the same way that our time is a subset
MicroPython added some extra stuff to it, which I was going to move to another module
tulip sleet
yes...
idle owl
I mean I'm ok with calling it adafruit_asyncio but if your intent was for the code to work on say an RPi with no changes, that might prove to be an issue.
It's about your intention, I suppose.
tulip sleet
do we intend that code with adafruit_requests work without changes on an RPi? Do we have projects that do that?
idle owl
Frankly, I would rather not have another library without adafruit_ on the beginning of it. NeoPixel is special.
idle owl
do we intend that code with `adafruit_requests` work without changes on an RPi? ...
No idea.
I can ask Brent, he might know.
But it obviously would need to be changed.
So...
tulip sleet
@slender iron what do you think about @lone axle 's idea of renaming asyncio to adafruit_asyncio, as with requests, hashlib, etc?
idle owl
Do you plan to, or did you already, publish it to PyPI?
tulip sleet
no, it will not be on Pypi, because you could just use the regular asyncio
did we publish adafruit_requests to pypi?
tulip sleet
yes, hmm
idle owl
Yes we do
tulip sleet
is requests strictly a subset?
solar whale
<@!343570430042308609> At least it does not seem to reset on the platforms i've ...
Good to know. I'll do some testing of it and supervisor.ticks_ms()
@timber mango what MCU are you testing it on?
idle owl
is requests strictly a subset?
I don't know it well enough in either form to have an answer to that.
tulip sleet
@idle owl - ok minor thing, if we did rename it, then I should capitalize the repo name, hunh? https://github.com/adafruit/Adafruit_CircuitPython_asyncio/ --> https://github.com/adafruit/Adafruit_CircuitPython_Asyncio/
slender iron
@tulip sleet I think my preference is asyncio because, as kattni points out, that is most compatible
idle owl
<@!330227457296957440> - ok minor thing, if we did rename it, then I should capi...
No. Keep it lowercase.
I prefer consistency across the libraries.
tulip sleet
in retrospect, we might have named adafruit_requests just requests
idle owl
We didn't. And it's on PyPI, so you can use it with Blinka. etc.
tulip sleet
The mpy-cross error message is something else :/ ~/Downloads/mpy-cross.static-amd64-linux-6.3.0 funcs.py dexpresed ing back (most recent call File "funcs.py", line 25stionusEUSyntaxError: altos ess not 'async for' or 'async with' outside async fu
idle owl
But, I don't know if it has things in it that requests does not.
tulip sleet
I think I will not change it now, but keep it under advisement. I think the bundle package prefix thing is something that should be fixed anyway. It makes the community bundle .yml kind of a mess right now. Not sure how I will fix the 6.x issue, though. This gets back to different versions per bundle, which we never solved
idle owl
Ok
timber mango
@solar whale That was the nRF
idle owl
add a `.bundle-6.x-disable` file ?? hmm ugh
I guess we can do it for this one library? With a comment about removing it when it's time, and removing whatever code you added to the bundle to make it work?
Or leave it in in case we need it again I guess.. I don't know.
I hope this is an unusual situation.
tulip sleet
so suppose we added, say, some other library that depends on 7.x features. Right now we conditionalize the code to do a version check
this is the first case where it simply won't compile on 6.x
i needto take a walk. This is quite a rathole.
idle owl
Ok. We'll be here.
idle owl
@tulip sleet Any clarity?
tulip sleet
no, I wasn't really puzzling over it, listened to a podcast instead. I have three PR's in the fire right now and am fixing another one. I am kind of inclined not to rename asyncio. I do think we need some kind of mild versioning on libraries for the bundle. I can't solve that otherwise, I think
idle owl
no, I wasn't really puzzling over it, listened to a podcast instead. I have thre...
I thought you'd already concluded you're not renaming it. I didn't think that one was still on the table. You could have it do both if it finds nothing, and only do the specified one if it finds the versioning file. Then you can add something only to asyncio, and we're not doing another massive lib sweep.
tulip sleet
the version file could be positive or negative. If positive it would declare a minimum version. It could be in the filename, or inside a file with a standard name
idle owl
the version file could be positive or negative. If positive it would declare a m...
Absolutely needs to be a standard filename. I don't want to mess around with mixed filenames.
That sort of thing makes patching really difficult.
@tulip sleet But you're saying it has to be present on all libraries for your idea to work?
tulip sleet
phone call, bbiab
tulip sleet
@idle owl if the file is absent, no restrictions. If the file is present, minimum version, <n>.x. How does that sound? Would you like it to be a dot-file?
I discussed the options at some length in an issue years ago
idle owl
<@!330227457296957440> if the file is absent, no restrictions. If the file is pr...
Yes, a dot file would be perfect.
tulip sleet
will find it
idle owl
And that sounds excellent.
manic glacierBOT
idle owl
<https://github.com/adafruit/Adafruit_CircuitPython_Bundle/issues/181>
Thanks. Let's do this.
tulip sleet
ok, I will figure something out!
idle owl
Feel free to tag me when ready for a look.
tulip sleet
none of the proposed solutions are a version file, it looks like
idle owl
Ok
crimson ferry
ESP32-S2 board with a SOLO module vs. with a WROVER module... surely I'm imagining that the SOLO is faster despite being essentially identical firmware on both boards? SOLO has integrated SPIRAM, WROVER has separate SPIRAM if I read the datasheets right, but the CP board files are the same with respect to speed etc.
idle owl
@languid whale Hello, do you mind if I DM you?
tulip sleet
@idle owl https://github.com/adafruit/Adafruit_CircuitPython_Bundle/issues/181 is really more about different versions of a single library for different versions of CircuitPython, rather than just "here's a new library that doesn't work with older versions of CPy".
so solving a somewhat different problem
idle owl
so solving a somewhat different problem
That's probably good since you had a different solution.
languid whale
<@!606613357491126293> Hello, do you mind if I DM you?
@idle owl please do!
dusk mauve
Is is possible to remove or exclude certain files from a frozen driver when building for a specific board?
Inspired by the BME280 library, with its basic and advanced classes, I'm thinking about suggesting an advanced driver revision for the APDS9960, since the driver gets pretty huge (as much as like ~12k vs ~4k) when exposing all of the config options.
Since APDS9960 gets frozen in for the prox trinkey builds I'd prefer to avoid breaking things there by excluding any prospective advanced file if possible.
slender iron
@dusk mauve my preference would be a separate library
dusk mauve
<@!253692424772124673> my preference would be a separate library
Thanks! That makes the most sense structurally. Would that be possible within the same repo? Or would it require setting up (and maintaining) separate repo altogether?
dusk mauve
Ahh, ok. Makes sense. Seems like a lot of work to enable features 99% of users probably wonβt care about. π
slender iron
all the more reason not to have everyone install it
plus separate repos are easier to find
dusk mauve
That makes a lot of sense.
Any examples of implementation of that model that come to mind? Iβll be taking a trip through some driver repos anyway so I may just stumble on one eventually
manic glacierBOT
solar whale
I've run in into an issue when trying to check for the existence of a module that will work on an MCU and on a Pi (with Blinka) . In many libraries, I see this ```try:
import foo
except ImportError:
import bar as foo
but
on raspberry pi, the failed import is a
ModuleNotFound error , not an ImportError
For my application, the tested module is not present on the Pi and I want to catch the error .
How would you reconcile this? If I just use a bare Except, pylint throws a fit...hmmm -- never-mind -- except ImportError appears to work on the Pi -- I'm confused, but relieved....
jaunty juniper
ah that would be because:
>>> issubclass(ModuleNotFoundError, ImportError)
True
Thanks -- I just stumbled across that myself ... I should have googled first...
@tulip sleet Do you know if the time.monotonic() loses accuracy on a Raspberry i in the same way it does on an MCU?
tulip sleet
<@!329766224093249548> Do you know if the time.monotonic() loses accuracy on a R...
does printing a float show more digits?
we are using internal object representation "A" for the RPi, which uses "boxed" floats, not immediate floats. I am honestly not sure if they are 32 or 64 bits. But they are at least 32.
e.g. how many digits does 1/3 show?
or time.monotonic()?
solar whale
nice! thanks -- I'll test a bit more later -- have to go to a meeting π¦
lone axle
hmmm -- never-mind -- `except ImportError` appears to work on the Pi -- I'm conf...
I don't recall the specific situation it came from but I just recently changed some code from catching a ModuleNotFound error to catching an ImportError I think there must be some situation out there where it does end up as a ModuleNotFound because I believe when it was originally coded that way it was tested successfully somehow.
Ah, still catching up. It being a subclass and thus ImportError working in both situations is good news.
onyx hinge
<@!329766224093249548> Do you know if the time.monotonic() loses accuracy on a R...
A 64-bit float still isn't unlimited precision. It looks like after a few thousand hours, nanosecond precision would be lost. after a few centuries, microsecond precision would be lost. (this is more of an issue for time.time(), where the epoch is 1970, not boot time or whatever) (I notice time_ns() has been added to cpython3 somewhere along the way, just like monotonic_ns())
solar whale
A 64-bit float still isn't unlimited precision. It looks like after a few thousa...
I think we can live with that! We are dealing with milliseconds
dusk mauve
Is there a core or 'standard' module/lib for CircuitPython that handles for RGB<->CCT<->Lux/Luminance type calculations? Looked for it in rainbowio, which seems to be a logical place for it to live someday, but so far its just the very useful colorwheel in there.
Its one of those things that seems pretty broadly useful but there are different ways to implement it, most leaning very heavily on floating-point math that could bog down non-FP-accelerated devices.
I'm curious about it since I'm wrapping up work on the APDS-9960 driver and it seems weird for it to include this colorutil utility module that has this broader functionality within it
jaunty juniper
Is there a core or 'standard' module/lib for CircuitPython that handles for `RGB...
have you looked into FancyLED ? https://learn.adafruit.com/fancyled-library-for-circuitpython
dusk mauve
have you looked into FancyLED ? <https://learn.adafruit.com/fancyled-library-for...
Not yet! Thanks for the link. π
jaunty juniper
I don't know if it has what you want though, but it has some color conversion stuff
dusk mauve
I don't know if it has what you want though, but it has some color conversion st...
there are actually some potentially useful things in there that could provide for some low-code examples of uses for the APDS-9960 sensor, at the very least. thanks again for the link. π Also looking at the adafruit_rgbled library but that looks quite a bit simpler
jaunty juniper
the rgbled library is a driver for driving PWM RGB LEDs
dusk mauve
Found some similar algorithms and patterns in the arduino/cpy libs for the somewhat similar TCS34725 sensor too
At first glance I really like the way its handled in the TCS34725 too.
Hah! Turns out the coefficients hard-coded in there come from one of the TCS families.. I love code archaeology, discovering the lineage of clever hacks through the ages.
slender iron
what do folks think about adding a regression issue label? that'd let us track how many issues are things that used to work
(and that testing could help prevent)
idle owl
Is there a core or 'standard' module/lib for CircuitPython that handles for `RGB...
rainbowio will likely remain simply the colorwheel due to size constraints.
idle owl
what do folks think about adding a `regression` issue label? that'd let us track...
Sounds like a good plan.
onyx hinge
@slender iron did you definitely need me/someone to take the december 27 discord meeting?
slender iron
I don't know yet. I haven't talked with becca about it
tulip sleet
i can do it
lone sandalBOT
slender iron
This is an interesting appnote about GPIO settings to reduce power consumption: https://www.st.com/content/ccc/resource/technical/document/application_note/group0/13/c0/f6/6c/29/3b/47/b3/DM00315319/files/DM00315319.pdf/jcr:content/translations/en.DM00315319.pdf
we tend to do floating inputs which are worse that pulling up or down
slender iron
@onyx hinge I think you can start on 7.x.x bugs next
I know deshipu would like to fix the displayio crash #5476
idle owl
@tulip sleet Those are the KB2040 files you emailed me right?
tulip sleet
yes, and I loaded the .uf2 on a KB2040 to test
idle wharf
I've not needed a debug build in a long time. Are they available on S3 or do I need to build my own ?
Got my Kaluga setup and going to get some logs for weird Sockets behavior as discussed in an issue on Requests repo.
crimson ferry
@idle wharf afaik you need to build your own, you'll need to disable some modules or something else to reduce the size https://github.com/adafruit/circuitpython/issues/4984
idle wharf
<@!624005998696333331> afaik you need to build your own, you'll need to disable ...
Thanks π
spiral elk
<@!624005998696333331> afaik you need to build your own, you'll need to disable ...
Would it also be possible to do custom partition magic on s2/s3 to avoid having to disable stuff?
crimson ferry
oh, probably, didn't think of that π
spiral elk
Since I think the second ota partition isn't used for anything currently
crimson ferry
who needs OTA when you're just debugging? π
in theory, I think there is OTA... I assume it works?
spiral elk
tinyuf2 just loads stuff into the first partition hardcoded IIRC.
crimson ferry
I think if you're really careful, you could write code on a CircuitPython device to update its own firmware and code.py
spiral elk
I think that if you're doing stuff where you need debug builds on a regular basis, it might be worthwhile to dedicate a board to a custom partition table.
idle wharf
I don't think (did not check) the ESP32S Debug builds won't fit on those boards. I built one and it loaded fine. Now to learn to add some print debugging to Socket.c and see if I can learn anything.
Because the test which ran a few times and failed on the UM Feathers2 fails on the second call on the Kaluga.
analog bridge
Just tested debug build on idf@v4.4, it seems to work now on boards with more than 4MB flash.
analog bridge
Ya the firmware overflows on 4MB boards because unlike normal builds the debug builds aren't optimized.
The initial dram0_0_seg overflow issue seems to have been fixed in the newer idf.
orchid basinBOT
orchid basinBOT
idle wharf
@slender iron not needing an answer right now, but I wanted to plant a seed for a potential deep dive topic. "How to Print debug in CP Core".
I don't know C, and I was poking around trying to use printf() just to see if I could anything and it didn't work, so I did some research and tried ESP_LOGD from the ESP-IDF, but still nothing coming out of the UART on my Kaluga. Anyway, I think with the searchable Deep Dive notes a 15 minute demo could be a great resource for everyone.
slender iron
mp_printf(&mp_plat_print, <regular printf>) will go out the regular serial
esp is special with it's ESP_LOG stuff
I think it's only connected up on debug builds
blissful pollen
mp_printf is engrained in my brain now unless I one day hook gdb up
slender iron
printf works for most things. gdb is nice for hangs and crashes
solar whale
@slender iron FYI -- File Glider has been updated by TestFlight --- PyLeap has not.
slender iron
ya, there is a review process that it has to go through to get in test flight
solar whale
No problem.
turbid radish
?serverinfo
digital shoreBOT
adafruit
lone axle
<@!252717193496756235> not needing an answer right now, but I wanted to plant a...
+1 for this I think it would be great to have the process documented somewhere as well. Either an "Advanced" learn guide, or maybe a section in the core readthedocs pages were the ones that came to mind for me. But really anywhere that makes it easy to find would be good.
dusk mauve
Finally finished my month-long educational project to improve the APDS-9960 driver and got the PR set up. Thanks to everyone that helped by answering my little questions along the way. And my apologies to whoever takes on the complex review. π
idle owl
@gilded cradle I have a WebSerial ESPTool question when you're around.
idle owl
Finally finished my month-long educational project to improve the APDS-9960 driv...
Congrats! And no worries. That's what the review team is for!
dusk mauve
Congrats! And no worries. That's what the review team is for!
I know, but I've been on the receiving end of some traumatic mega-PRs (time-constrained ones at that, lol) so I do try to be considerate π
idle owl
I know, but I've been on the receiving end of some traumatic mega-PRs (time-cons...
Appreciated π
dusk mauve
I did learn how to use pre-commit more effectively, as well as preview doc changes quickly by using sphinx to build the doc locally. That was useful. π
idle owl
I did learn how to use `pre-commit` more effectively, as well as preview doc cha...
Great! Sphinx can be awful, so that's a good learning experience.
dusk mauve
Great! Sphinx can be awful, so that's a good learning experience.
I can imagine the frustration it can cause, given how dang powerful it is. And how arcane some of its syntax is. But I was really happy with how well the docstring auto-formatting worked out, after some trial/error of course
idle owl
I can imagine the frustration it can cause, given how dang powerful it is. And h...
That kind of sums up figuring out Sphinx π
It has taken two and a half years to get to a point where I can pretty much streamline my way through issues. Because I've finally seen them all and slowly remember what the solutions were.
ember iris
Sphinx is so powerful that you have to fight against it to do something simple in my experience. But it's so powerful it's worth the headache
dusk mauve
It has taken two and a half years to get to a point where I can pretty much stre...
only because that "slowly remembering" has always been super slow for me. π
ember iris
I keep cheat sheets everywhere! It really helps until you lose them π
I need a cheat sheet to finding my cheat sheets...
dusk mauve
Oh man I know that feel Keith π₯²
idle owl
It also helps that I often asked on Discord and Discord has a great search feature π
gilded cradle
<@!478616919721771019> I have a WebSerial ESPTool question when you're around.
Sure, I'm back.
idle owl
Sure, I'm back.
Ok, so we have this step in the explanation for using the WebSerial ESPTool for enabling it in old Chrome. I looked it up, the last version that has webserial enabled by default was released in March 2021. Do we need to keep that step in there anymore? Or should we assume folks have updated Chrome since March?
gilded cradle
Ok, so we have this step in the explanation for using the WebSerial ESPTool for ...
I think we left it in case anybody is running something like Linux which doesn't auto update and has an old version. Maybe we can change it to say that it only applies to older versions.
idle owl
I'm thinking adding an alert element that links to a section at the end of the page that explains it so it's not the first step.
gilded cradle
Ok, that works
idle owl
It already says only older versions
But it seem weird having it first in the list of steps
gilded cradle
Yeah
idle owl
Yeah
Thanks Melissa!
Actually, what I'd really like to know is.... what error happens if you try to use it without WebSerial enabled?
I need to figure out if I can install an old version of Chrome.
And not break everything, lol.
gilded cradle
Actually, what I'd really like to know is.... what error happens if you try to u...
It will show you a message that you need to enable it and not work otherwise.
idle owl
It will show you a message that you need to enable it and not work otherwise.
OK, I'm going to try to get a screenshot of that message for this section. If I can figure this out.
gilded cradle
OK, I'm going to try to get a screenshot of that message for this section. If I ...
Ok, sounds good. If you can't, I can try tricking it and get you a screenshot.
idle owl
Ok, sounds good. If you can't, I can try tricking it and get you a screenshot.
OK, I'll try this for a short while and then ask for help if I can't get it going.
solar whale
has anyone tried the broadcom build for the Raspberry Pi 4B? I have built it for both the Raspberry Pi Zero 2W and the 4B.The Zero 2W boots fine, but the 4B is not booting. I have a USB-Serial connected to TX,RX, GND -- On the 2W it boots to the REPL, On the 4B I see no activity at all...
idle owl
Ok, sounds good. If you can't, I can try tricking it and get you a screenshot.
Managed to get it!
gilded cradle
Awesome
slender iron
@solar whale nothing even on an hdmi display? the bootloader should show something there
solar whale
I don't have an HDMI connected now -- Will try that...
It just goes to the multicolored square
slender iron
hrm, that is into CP
I wonder if it is USB related
I don't usually have usb connected immediately
solar whale
OK -- that will be a good learning experience π Does the size of the SD card Matter -- I have 8G on the 2W but 32G on the 4
manic glacierBOT
slender iron
it shouldn't but I haven't exhaustively tested it
solar whale
OK -- I'll explore -- tomorrow -- thanks.
slender iron
thanks!
solar whale
I was mostly curious if you had tried the 4B in addition to the CM4
solar whale
I feel so cutting edge π
slender iron
you always are π
solar whale
sorta like my first run down a "black diamond" ski trail -- terrifying...
hidden oxide
I need to figure out if I can install an old version of Chrome.
VM OS??
hidden oxide
Aha
idle owl
It's bonkers complicated to find older Chrome installers through the Google site.
There are tons of other sites that have them, but if you want them from Google, you have to do some steps.
hidden oxide
VMWare Workstation with my pentesting OS and Windows 10 and Raspbian is really hard on a 4GB RAM computer
That's my setup
idle owl
Oi.... yeah. 4GB with VMs sounds like a stretch.
hidden oxide
And 3 OS's
Especially my pentesting OS struggles - what with brute force numbers and Hashcat
and Wifi attacks too
I upped my wifi security after I hacked it with no info to the system within 30 seconds
Makes you wonder
hidden oxide
There are tons of other sites that have them, but if you want them from Google, ...
I bet
I've got Mac Maverick - I don't use it anymore for that reason
idle owl
@gilded cradle Another WebSerial question if you're around.
@gilded cradle Here's the question, feel free to answer whenever you get a chance.
For the purposes of loading the bootloader, this template page will not refer to esptool. Only WebSerial ESPTool. Do you really need to identify the serial port first and separately if you're going to use WebSerial ESPTool anyway? i.e. Do I need to walk people through finding the COM/etc port before telling them to use WebSerial. Because WebSerial will display what it finds anyway, won't it?
gilded cradle
<@!478616919721771019> Here's the question, feel free to answer whenever you get...
If I recall correctly, it displays a list of ports it finds.
idle owl
Right, exactly.
So I'm thinking using Terminal or Device Manager to find the port is pointless.
But I wanted a second opinion.
gilded cradle
So I'm thinking using Terminal or Device Manager to find the port is pointless.
Yeah, kinda redundant.
idle owl
I can basically say "Only plug in the board you want to flash" and then walk through the ESPTool steps.
Ok
Thanks!
gilded cradle
Sounds good
idle owl
@gilded cradle Blergh. So.... I got this error after erasing using WebSerial ESPTool ..... Did I get the board in a state where I need to use esptool to flash it or something? I erased it, disconnected, and am now trying simply to reconnect, and it gives me this instead: Error: Software loader is resident at 0x40028000-0x40028FE8. Can't load binary at overlapping address range 0x40028000-0x40028FE8. Try changing the binary loading address.
Sorry to keep bothering you, I have it in my mind that you dealt a lot with WebSerial... maybe I'm wrong about that, and if so, let me know...
gilded cradle
<@!478616919721771019> Blergh. So.... I got this error after erasing using WebSe...
Yeah, I'm not sure what's up with that, but I think there's an issue open for it.
idle owl
Ahhh ok.
I guess I'll try to get esptool going again. I used it a long time ago, iirc.
Thanks
gilded cradle
yw
@idle owl can you add notes to https://github.com/adafruit/Adafruit_WebSerial_ESPTool/issues/17 to help with figuring out the steps to reproduce the error?
idle owl
Absolutely.
gilded cradle
Thanks
idle owl
Updated.
idle owl
@gilded cradle I managed to recover the board with esptool and so of course I immediately followed the same steps as earlier on WebSerial ESPTool, and same results. Updating the issue to mention I reproduced it as well.
gilded cradle
<@!478616919721771019> I managed to recover the board with `esptool` and so of c...
Awesome. Thank you.
manic glacierBOT
manic glacierBOT
viscid pine
I'm trying to understand why the rp2040 ports specify external flash
while it is external to the main chip, isn't it really internal in circuitpython terms?
never mind i see now, it's used to gen stage2
manic glacierBOT
manic glacierBOT
tulip sleet
@idle owl yay on Factory-Reset-Images
idle owl
<@!330227457296957440> yay on Factory-Reset-Images
π I don't know how quickly it's going to fill up, but we'll see. I needed somewhere to put the bin files for this ESP32-S2 template.
@tulip sleet So you sent me a .ino and a .uf2. I assume they are identical, except one you can drag and drop to RPI-RP2, and the other you need Arduino to load?
tulip sleet
yes, I actually compiled the .ino into the .uf2. I think adding both fine would be fine.
idle owl
Ehhh....
The NeoPixel one is a default NeoPixel example in Arduino. I'm not sure I want to deal with .ino files in this repo.
Maybe.
tulip sleet
The NeoPixel pin number may vary between different RP2040 boards?
idle owl
The NeoPixel pin number may vary between different RP2040 boards?
Oh hmm right. Wait... no. Hold on let me check. She started adding PIN_NEOPIXEL to everything.
Looking to see what's in this.
tulip sleet
that .ino may have been written before KB2040 got added to Arduino
idle owl
Noop. PIN_NEOPIXEL not used.
OK. I guess I'll include both.
I'm not giving it its own directory though. User can do that! :shakes fist:
tulip sleet
oh, sure, most people will just use the .uf2 anyway
idle owl
<@!330227457296957440> yay on Factory-Reset-Images
We're going to add them to the PCB repos instead. And only UF2/bin. Deleted the factory-reset-images repo per Limor.
idle owl
@lone axle I have need of your assistance. I need to update the PCB cookiecutter to create a new folder with a dynamically generated README.... and I'm confused about how to do that. Oh hah. I think I figured it out. Nevermind. π
lone axle
<@!382939733107408897> I have need of your assistance. I need to update the PCB ...
Happy to help if it turns out you need still.
idle owl
Happy to help if it turns out you need still.
Thanks! Appreciate it. But it looks like I did it right. We'll find out the next time I need to use it! π
@lone axle wait I lied. it's way more complicated than what I did.
I do need help, because it needs to ask "Is this a dev board? Y/N", and then only include the new folder if Y.
I thought I was so slick there. But... I realised it's not quite what we need.
lone axle
I do need help, because it needs to ask "Is this a dev board? Y/N", and then onl...
This is the right repo that your adding it to right? https://github.com/adafruit/cookiecutter-adafruit-pcb
idle owl
Yes.
I could try to figure out how it's done in the other cookiecutter..... but I don't know how to generate a folder based on the answer to a question....
lone axle
Okay, to add the extra question it would be a new entry in this file: https://github.com/adafruit/cookiecutter-adafruit-pcb/blob/main/cookiecutter.json
idle owl
OK, I'm there
idle owl
Okay, to add the extra question it would be a new entry in this file: https://gi...
Added "Is this a microcontroller board?" : "N", to the beginning.
and like in the lib cookiecutter, I'd like "y" and "yes", "n" and "no" to both work. Because I never remember π
lone axle
perfect. that should make it ask that question and default to "N" I believe. Then in order to use the answer to that question you'd work in some logic into the weird looking {{ template_var }} or {% template_logic %} tags.
idle owl
in the directory name?
lone axle
yep
So if factory-reset directory existence needs to be conditional on that question I think the name of that directory should be something like:
{% if cookiecutter.dev_board %}factory-reset{% endif %}
{% if cookiecutter.dev_board in ['y', 'yes'] %}factory-reset{% endif %}
lone axle
However I do think you'll need to change the entry in cookiecutter.json a bit. The key for that dictionary item should probably not have spaces. I would make the key something like dev_board and then the value is the text of your question I believe.
idle owl
Hmm ok
Documentation I guess!
ok
updated
So, do I have to conditionalise the files within the directory or because the directory is conditional, it "just works"?
lone axle
I'm not 100% certain, but I think the directory conditional would carry over to them.
idle owl
Ok, testing.
Now it's failing saying the directory already exists.
Which it does not.
Like it's trying to create it twice or something
The top level directory
not the factory-reset directory
lone axle
Poking a little bit on my end as well. It seems like the values in that dictionary are the default values, not the text of the question. Beyond showing the values of the default I'm not sure if there is a way to specify different text for the question.
idle owl
That's fine, I can document what it means, and really only Dylan and I are doing it that I'm aware of. Maybe Limor.
Can update README.
But it's not running now. Which is a problem.
pre-generate
5000 is PID
Error: "/Users/kattni/repos/cookiecutter-adafruit-pcb/Feather_ESP_pcb" directory already exists``` is how far it gets.I can tinker with it for a minute. I've never actually done a conditional directory but I have a vague idea of how I think it will work. Don't want to send you too far down the rabbit hole.
idle owl
Ok, thank you.
idle owl
Without that folder, it runs properly.
lone axle
I think there may be a difference in what is needed for a conditional directory vs. and conditional file.
for files if the name equates to empty it just doesn't get created. But for directories if the name equates to empty it tries to do something else more weird. I think post_gen hooks are the way to make it work though. It runs those after creating things. So factory-reset can get created by default and then deleted if uneeded from the post_gen. Working out the specifics of that now.
idle owl
Oh ok