https://docs.python.org/3.8/library/os.html#os.confstr we could implement unix-compatible os.confstr for more information
#circuitpython-dev
onyx hinge
os.confstr(os.confstr_names['_SC_CIRCUITPYTHON_BOARD']) -> 'adafruit_feather_rp2040' or whatever
tulip sleet
https://docs.python.org/3.8/library/os.html#os.confstr we could implement unix-c...
This seems really specific and unusual:
>>> os.confstr_names
{'CS_GNU_LIBC_VERSION': 2, 'CS_GNU_LIBPTHREAD_VERSION': 3, 'CS_LFS64_CFLAGS': 1004, 'CS_LFS64_LDFLAGS': 1005, 'CS_LFS64_LIBS': 1006, 'CS_LFS64_LINTFLAGS': 1007, 'CS_LFS_CFLAGS': 1000, 'CS_LFS_LDFLAGS': 1001, 'CS_LFS_LIBS': 1002, 'CS_LFS_LINTFLAGS': 1003, 'CS_PATH': 0, 'CS_XBS5_ILP32_OFF32_CFLAGS': 1100, 'CS_XBS5_ILP32_OFF32_LDFLAGS': 1101, 'CS_XBS5_ILP32_OFF32_LIBS': 1102, 'CS_XBS5_ILP32_OFF32_LINTFLAGS': 1103, 'CS_XBS5_ILP32_OFFBIG_CFLAGS': 1104, 'CS_XBS5_ILP32_OFFBIG_LDFLAGS': 1105, 'CS_XBS5_ILP32_OFFBIG_LIBS': 1106, 'CS_XBS5_ILP32_OFFBIG_LINTFLAGS': 1107, 'CS_XBS5_LP64_OFF64_CFLAGS': 1108, 'CS_XBS5_LP64_OFF64_LDFLAGS': 1109, 'CS_XBS5_LP64_OFF64_LIBS': 1110, 'CS_XBS5_LP64_OFF64_LINTFLAGS': 1111, 'CS_XBS5_LPBIG_OFFBIG_CFLAGS': 1112, 'CS_XBS5_LPBIG_OFFBIG_LDFLAGS': 1113, 'CS_XBS5_LPBIG_OFFBIG_LIBS': 1114, 'CS_XBS5_LPBIG_OFFBIG_LINTFLAGS': 1115}
and inconvenient
tulip sleet
and I don't see an analogue in the host computer world
how are you feeling about board.id?
onyx hinge
I think the main "problem" with board.id is requiring edits to all the board.c files.
I could try to automate it via awk
jaunty juniper
interestingly it's hardcoded in every port without using MICROPY_HW_MCU_NAME but it seems to have the same value
onyx hinge
I/somebody
tulip sleet
well, the PR has already done it, we just need a sed over it
jaunty juniper
I did automate it with python to begin with
onyx hinge
oh there's a PR? even better.
tulip sleet
i think a no-parse solution is the easiest for the libraries and code that need to check
jaunty juniper
why do some boards use board_global_dict_table and others board_module_globals_table ?
tulip sleet
I think there is no good reason, it was an historical accident
I like the latter better. That could be another global sed fix.
the name used in shared bindings for the module dicts is, for example, STATIC MP_DEFINE_CONST_DICT(busio_module_globals, busio_module_globals_table);
actaully it's not so consistent in modules either
well, dict_table is used in class dictionaries, but globals_table is used in module dictionaries, it looks like
since board is a module dict, globals_table follows that pattern
onyx hinge
I wonder if we could have a macro at the top of the board globals_table that would expand to zero or more entries in the globals table .. it could reduce duplication, such as by automatically adding board.I2C / board.SPI, and of course this new board.id
onyx hinge
STATIC const mp_rom_map_elem_t board_global_dict_table[] = {
+ CIRCUITPYTHON_BOARD_STANDARD_ITEMS
{ MP_ROM_QSTR(MP_QSTR_D10), MP_ROM_PTR(&pin_GPIO10) },
```something like thatI was thinking maybe replacing the whole board_module_globals_table and board_global_dict_table line with a macro
CIRCUITPYTHON_BOARD_START_PINSC
...
CIRCUITPYTHON_BOARD_END_PINSC
there are boards that only have I2C, etc.
onyx hinge
hiding {} inside of macros doesn't make me thrilled 😕
@tulip sleet the STANDARD_ITEMS macro can be carefully assembled so that it works if there's no SPI, etc
tulip sleet
MicroPython generates the pin dict from a .csv file. I took that out because it was harder to read than the expanded C code
onyx hinge
but this is making work where none is really needed atm
tulip sleet
but this is making work where none is really needed atm
agree
One reason to go back to a machine-readable list is to generate the IGNORE settings automatically (or turn them into INCLUDE settings). Scott and I discussed that briefly, I can't remember in what setting.
but the current list could be parsed in an ad hoc way too
jaunty juniper
how about a toml file, would that be more readable ?
tulip sleet
oh yeah, and it could have comments, which is the big thing I missed in the csv file
jaunty juniper
[pins]
TOUCH2 = pin_PA03
NEOPIXEL = pin_PA15
something like that
without even the [pins] section actually
tulip sleet
that info could also be used to make mpconfigboard.h, or at least part of it
i think we need a [default_devices] section.
Ultimately we want a board and its details to be described in toml, whatever those details are: name, id, pin names, crystal or not, etc.
modules to include, etc.
whatever is in mpconfigboard.mk and mpconfigboard.h
the q is how much easier would it be to do this than the current scheme, so how worth it is it?
we had several errors recently involing the IGNORE pins, so that's a slight motivation
Information about a board could also include info that would be in circuitpython-org, so more of that could be added automatically (though adding the pics is an issue)
lone axle
On PWMOut there was an argument in the constructor variable_frequency https://circuitpython.readthedocs.io/en/latest/shared-bindings/pwmio/index.html#pwmio.PWMOut is there something analogous for for that in PuseOut now that does not need the PWMOut object? I see that frequency and duty_cycle arguments do exist for PulseOut but don't see anything listed in the docs for variable_frequency https://circuitpython.readthedocs.io/en/latest/shared-bindings/pulseio/index.html#pulseio.PulseOut
slender iron
@onyx hinge are you looking at the esp wdt crash?
onyx hinge
@slender iron no just briefly earlier
slender iron
the long lived thing is a good lead
we must be handing a pointer to the idf that we move
onyx hinge
@slender iron yeah I re-tested in the hopes that the one long-lived problem we'd fixed would fix that too, because it seemed like a really good lead and nothing I'd looked at last week. when that didn't work, I reasoned that I could pretty quickly determine if the long-lived moving process was still in the chain of events, and it looks like it is.
slender iron
I hadn't thought of that yet because the i2c display move is different from the long lived one. The display is moving out of the heap to supervisor allocated space
I just thought it might be the scanning structure but it's not...
idle owl
@tulip sleet Are there any boards that do not have adafruit_bus_device natively?
The CP module version.
Oh it looks like a bunch don't.
Nevermind. I was going to see about having the Project Bundler ignore it, but that's not an option it looks like.
idle owl
@gilded cradle Is zip file manipulation a native thing in Linux? Or specifically, on Raspberry Pi?
idle owl
@tidal kiln Yeah, trying to figure out once and for all how to structure breakout guides with the Project Bundle.
Because it used to be "CP Install" "Python Install" "CP and Python usage"
but it doesn't really work that way anymore with the Bundler.
I think I'll leave it as I have it now, and assume that folks who are using this stuff on RPi etc know how to create a file and run it.
tidal kiln
since they'll end up with a .zip from the bundler?
idle owl
I'm not indicating Python folks should download the zip.
I thought I might for a minute there.
Then decided against it.
But maybe that's a better idea?
Yeah, that was why I asked in the first place though.
tidal kiln
yah, on a pi you wouldn't really need anything but the code, right?
idle owl
Right.
tidal kiln
since you'd pip install the libs
tidal kiln
so could be confusing...i downloaded the .zip and what do i do with all these files?
jaunty juniper
(and the libs in lib/ would not be seen anyway)
idle owl
I thought about including a paragraph stating something like "Save the following as code.py where it's convenient, and then navigate to it via command line and run python3 code.py"
Honestly, if we weren't supporting 6 and 7, I'd have them download the bundle.
Because the first level would contain code.py
jaunty juniper
what about projects that have images and support files ?
idle owl
but it doesn't, so it's more confusing.
Oh good point.
Ugh.
Most breakout guides don't have that though.
tidal kiln
this is for breakout guides?
idle owl
Like, this is a simpletest for a CO2 sensor.
jaunty juniper
ah breakouts
tidal kiln
i just wget it, but not sure how easy that'd be to guide-ify
tidal kiln
a bit clunky, but can use the "copy code" to get it into some text editor and save
idle owl
i just wget it, but not sure how easy that'd be to guide-ify
Could include this in the library installation section
like "Library installation and Example download" or something
What we used to do was use the REPL for the example demonstration
but A. doesn't work with all examples and b. doesn't allow for Bundlefly downloads.
So it's not a great solution, and we haven't addressed how to handle the Python side of that.
Or add a separate Python Example Download section below the lib install section might work better.
idle owl
i just wget it, but not sure how easy that'd be to guide-ify
So this would be wget https://www.github.com/url-to-simpletest ?
tidal kiln
it'd need to be the "raw" url
idle owl
or go ahead and get bundle zip but only grab code.py from it?
Yeah but it's buried in a CircuitPython 6.x/7.x folder
For now.
tidal kiln
which is where everyone seems to go wrong and end up with HTTP instead of python code
tidal kiln
tidal kiln
wget the second one and you get http
idle owl
When I click "raw"
tidal kiln
yep. or that url.
idle owl
Ok.
@tidal kiln What do you feel would create less support issues? A section telling them to dig code.py out of the bundle, or a short section with the wget command listed?
tidal kiln
dig
idle owl
Ok. Will do.
tidal kiln
could give them an unzip snippet to just extract code.py
idle owl
@tidal kiln How's this? I haven't added the wiring diagrams yet, but take a look at the rest. https://learn.adafruit.com/adafruit-scd-40-and-scd-41/python-circuitpython?preview_token=pfsUJ0diT27QVNKcNzw1LQ
tidal kiln
yah. thats probably ok.
there's a separate subfolder for CP versions, right? in the zip
idle owl
there's a separate subfolder for CP versions, right? in the zip
Yes
Mention it doesn't matter which version you choose?
idle owl
Ok
tidal kiln
right. yah. just something that mentions it'll be buried, and doesn't matter which.
idle owl
Trying to figure out how to keep it agnostic because once 7 is stable, we don't need 6 support in the bundle.
I don't really want to go back and update every guide.
At least I think that is the plan.
"Extract the resulting zip file, and navigate to the code.py file. The CircuitPython version does not matter."
then if there's no version, shrug.
Not certain I like it. But it's a start.
tidal kiln
@idle owl unzip -o -j FooBundle.zip */*/code.py
idle owl
Woah what does that do
tidal kiln
-o to force overwriting existing file, -j to "junk" directory structure (no subfolders get created)
and then wildcards to go down to subfolder
tidal kiln
yep
idle owl
Dang. Nicely done.
tidal kiln
a bit hacky, it'll grab all versions
and overwrite as it goes
but should all be same, so doesn't matter
tidal kiln
$ unzip -j -o Macropad_Hotkeys.zip */*/code.py
Archive: Macropad_Hotkeys.zip
inflating: code.py
inflating: code.py
Oooh, even better, the library examples are always in a folder called examples.
I mean MacOS appends a number to them
but ideally it'll be what it is for folks on Linux.
idle owl
Fair enough
tidal kiln
let me actually verify that's all same on pi
tidal kiln
that was tested locally
jaunty juniper
I'm getting errors 500 downloading bundles
idle owl
@jaunty juniper Really?
tidal kiln
which guide? i just grabbed that macropad hotkeys bundle ok
jaunty juniper
ah it came back
tidal kiln
@idle owl yah, looks same on pi
idle owl
<@!330227457296957440> yah, looks same on pi
Perfect.
jaunty juniper
it was there https://learn.adafruit.com/touch-deck-diy-tft-customized-control-pad/code-the-touch-deck where the link gave me 500: https://learn.adafruit.com/pages/21870/elements/3086140/download?type=zip
wait it looks like every time it's the first time the link gives me 500
or it's random
idle owl
I chose the Rubber Ducky guide.
tidal kiln
just downloaded that touch deck bundle ok
idle owl
I was in a private browser window, not logged in.
timber mango
Does anyone know if the nRF52840 Feather CP port initialize busio.SPI differently when the device is running on battery, rather then on USB power?
I'm having a very strange problem with running a SPI device on it, when on battery power via JST connector, and i've narrowed it down to being a matter of what the nRF is running on when the busio.SPI class is instanced.
Bit more info here: https://forums.adafruit.com/viewtopic.php?f=57&t=182604&p=887310
idle owl
ah it came back
Turns out it's an issue. Justin's looking into it.
@tidal kiln Anne thinks it might be confusing to microcontroller folks, having all the CPython instructions in the middle. I'm not sure how else to do it. I guess I'll run it by Limor and see if she wants it done some other way.
timber mango
Does anyone know if the nRF52840 Feather CP port initialize busio.SPI differentl...
I actually just worked around the issue, by initializing a dummy SPI instance, and then another one, which is then used.
Looking at https://github.com/adafruit/circuitpython/blob/main/ports/nrf/common-hal/busio/SPI.c i can see that the code will prefer SPIM3 first, which is 32Mhz, and then SPIM2 which is 8Mhz. So guessing there is some issue with initializing SPIM3 when running on battery?
Or maybe the frequency is being set incorrectly.. Unfortunately i don't have any equipment to verify this
gilded cradle
<@!478616919721771019> Is zip file manipulation a native thing in Linux? Or spec...
Glad @tidal kiln was able to help. I wasn't sure of the answer.
idle owl
Glad <@!337989520962093057> was able to help. I wasn't sure of the answer.
Ah right on. Thanks for following up anyway!
solar whale
@tulip sleet just an FYI - you might find this interesting given prior struggles you have had with SPI on the nRF52840 https://forums.adafruit.com/viewtopic.php?f=57&t=182604&p=887344#p887344
timber mango
<@!329766224093249548> just an FYI - you might find this interesting given prior...
I wonder if its related to the hardware errata on the SPIM3 interface. Seems there is already some workaround implemented in CP.
solar whale
Very possibly -- I vaguely recall the issue and need to revisit it. Thanks for all your effor tracking this down
timber mango
Its all very weird how its related to running of a battery though
solar whale
Sure is...
I have been using batteries on M4 and esp32s2 boards regularly, guess I never tried an nRF52840
timber mango
Thanks for looking into it!
I was surprised how low the sleep power usage was on the nRF, so wanted to switch to this from RP2040 which is not great for low power sleep
its like 0.3ma in time.sleep on nRF
solar whale
I have been using an esp32s2 (unexpected maker feathers2) for sleep with the rfm9x -- it has been working well.I don't have the numbers ofhand.
I'll be interested to try the nRF as well now that you have prodded me 😉
slender iron
@tulip sleet do you want me to do a beta.1?
tulip sleet
do you want to wait for some more PR's before that? I would have time to do one on Friday.
slender iron
@tulip sleet maybe? I was thinking the HID fix would be a bit urgent
tulip sleet
I put a note in the guide to use Absolute Newest on or after a certain commit.
also i well might have time to do a release starting tonight and finishing tomw morning
slender iron
idle owl
@ember iris SCD-30 has been running on RPi for 57 hours or something. I'm going to stop it at this point. I clearly couldn't reproduce it.
ember iris
<@!459119350851567626> SCD-30 has been running on RPi for 57 hours or something....
Yeah after they said it was a bread board my thought was that the real issue is a connection somewhere. Clock smearing seems like a really niche bug, and if you can't reproduce it I think it might not be the root cause
slender iron
what does purple blink on magtag mean?
onyx hinge
@slender iron I think it's from tinyuf2
ports/esp32s2/README.md:-
PIN_BUTTON_UF2is gnd when 2nd stage bootloader indicator is on e.g RGB led = Purple. Note: since most ESP32S2 board implementGPIO0as button for 1st stage ROM bootloader, it can be used for dual-purpose button here as well. The difference is the pressing order:
gloomy shuttle
it blinked purple for me for some reason
Mine does that when using code that causes the issue #5021 - https://github.com/adafruit/circuitpython/issues/5021
onyx hinge
@slender iron did you verify my finding that effectively disabling the "make long lived" machinery fixed it?
onyx hinge
true
onyx hinge
I wonder if @gloomy shuttle 's reduced test case will run on kaluga so we can get a debugger on it
looks like should
@slender iron the long-lived object is a copy so what if the object has a finaliser? will it run earlier than it should be permitted to?
.. when the original is gc'd
slender iron
ya, maybe! I hadn't thought about double running a finaliser
onyx hinge
that said I think one of the permutations I tried earlier was supposed to skip moving when a thing had a finaliser
none of the objects passed to gc_make_long_lived in the failing case seem to have a finaliser
onyx hinge
that makes no sense, so goodnight
solar whale
@timber mango I opened an issue to track the nRF52840/rfm9x battery power issue. https://github.com/adafruit/circuitpython/issues/5233
lone axle
Is it possible to change the frequency of a PulseOut after it's created? It's done using the old API in this project here: https://github.com/adafruit/Adafruit_Learning_System_Guides/blob/c7471f8fcf2b77c619431c87025ac3df447cdc02/CircuitPython_TVBGone/cpx_main.py#L67 Is there some way to achieve the same functionality using the new API for PulseOut that accepts the pin rather than the PWMOut object?
tulip sleet
Is it possible to change the frequency of a PulseOut after it's created? It's do...
That is a deficiency of the new API; we should add a PulseOut.frequency. Thanks for noticing that.
idle owl
Yeah after they said it was a bread board my thought was that the real issue is ...
I hadn't followed up on the forum post. Good to know. Regardless, it means I don't have to update the guide.
onyx hinge
well this is a new one. it looks like you have to talk on an I2C bus to turn on the neopixel power
idle owl
On what
onyx hinge
at power on PERI_POWER_ON is floating. I don't know my mosfet symbols well enough to know if that will pass power through Q12 or block it
@idle owl this espressif "hmi devkit"
idle owl
Ah
onyx hinge
PERI_PWR_ON is on a TCA 9554 I/O expander
idle owl
Huh.
onyx hinge
@tulip sleet do you read schematics well enough to tell me if PER_VCC_3V3 is on if PERI_PWR_ON floats?
V_GS is 0v so I imagine that means it'll be off
tulip sleet
it is a P-channel, so it's on when the gate is grounded. This is pulled up, so ground to turn off
so it looks like it's normally off
floating is not great
CORRECTED!
onyx hinge
well, I mean, it's not floating because there's the pull
tulip sleet
wait a minute, I might be wrong
onyx hinge
when -VGS is zero, it is off, as -VGS gets bigger it conducts.
while N-channel work with +VGS
.. right?
tulip sleet
also need to consdier depletion vs enhancement
onyx hinge
it's starting to look like help-with-electronics in here 🙂 thanks dan
tulip sleet
sorry, i have not used MOSFETs much and I have to look things up
idle owl
It was help-with-electronics in here the other day with the SPEAKER_ENABLE pin checks because I needed Dan's help 😄
onyx hinge
I propose that henceforth they have to put "this way up" stickers on each transistor
tulip sleet
@onyx hinge p-channel depletion are rare
idle owl
Anyone mind taking a look: https://github.com/adafruit/Adafruit_Learning_System_Guides/pull/1807
idle owl
Forgot how much I don't like working with Pico. This counting-the-pins-to-find-the-one-you-want-because-there's-no-labels thing gets old really quickly.
Makes me pick GP0 for everything I can.
"But the ground pins are square!" Yeah ok, I soldered headers on it, they all look like solder-blobs now, thank you.
@onyx hinge As soon as I get the code embedded into this page, I'll send it your way to take a look at it. It's the MP3 template filled in for Pico.
onyx hinge
hmmm the factory test app comes up scrambled on the hmi devkit, I wonder if the app is busted or the device
timber mango
<@456226577798135808> I opened an issue to track the nRF52840/rfm9x battery powe...
@solar whale Thank you - I was unsure whether to create an issue on it, as i don't know if its a hardware issue or a CP issue.. But i guess it might be both
onyx hinge
idle owl
For some reason, I consistently have about a 50% chance I will type "decoder" or "encoder" as "decode" or "encode". Writing guides for MP3 and rotary encoders gets frustrating. 😄
onyx hinge
it's all sandwiched together
idle owl
I thought as much. So much for that plan.
onyx hinge
I'll probably grab & build their default app for myself, I have experienced that often devices ship with software that isn't quite right
but I'll also open it up at some point, I have to if I want to install the battery
idle owl
Ah
onyx hinge
Adafruit CircuitPython 7.0.0-beta.0-44-gd8b3d5af0-dirty on 2021-08-26; HMI-DevKit-1.1 with ESP32S2 ok that's a good start 🙂
idle owl
@onyx hinge Here's the preview of the MP3 template, filled in this time, for Pico. Can you give it a read-through and tell me if I need to / should change anything? https://learn.adafruit.com/mp3-playback-rp2040/pico-mp3?preview_token=wRJvd7RoFcwFRd1Dfr9KLg
onyx hinge
should it say "following additional hardware"?
You'll need the following hardware to complete the examples on this page.
onyx hinge
should there be a spot to explain which pins can be used? for instance on rp2040 it can be any pin, but on samd51 it can be just one of 2 pins if I recall correctly...
or is giving an example of SOME usable pin good
idle owl
Hmm.
That's valid.
I can add something in about that.
I think it's worth including.
We do it for other concepts.
I2C, PWM, etc.
onyx hinge
the project zip is not including any mp3 files I think
onyx hinge
this is wrapped weirdly for me in learn, 1920x monitor, default zoom level
idle owl
hmm. That's Learn wrapping it weird. But I'll fix it in the file.
onyx hinge
I think that concludes my feedback 🙂 it's nice and clear overall but I found some nits to pick 🙂
idle owl
This is why I asked you to take a look before asking for moderation on it. I appreciate it.
onyx hinge
you're welcome
idle owl
@onyx hinge How do you know what pins are valid? For I2C, I wrote a script that tries to create the I2C object on each pin, but MP3 isn't constructed like that.
idle owl
Oh. SAMD51 is limited for AudioOut then?
onyx hinge
yeah I thiink only A0/A1
ornate breach
What was that thing Ladyada did for generating board pin reference printouts?
idle owl
You mean PrettyPins?
ornate breach
I can’t remember it for the life of me
Was that the one she covered on desk of Ladyada?
idle owl
Do you mean the graphical pinout diagrams?
ornate breach
It created like a pin reference for svg board images?
idle owl
I'm the one who does most of those 😄 But she and PaintYourDragon wrote the software to do it. We called it PrettyPins, and it stuck.
Yes, PrettyPins.
It can be finicky, and requires a CSV of possible pins for a particular microcontroller, and begins with a Fritzing object. (which is a zip of SVGs, really).
prime flower
Is there code anywhere which shows which folders/files are created on a default CIRCUITPY filesystem, after it's been flashed?
onyx hinge
@prime flower yeah let me find you a link
prime flower
thx, rewriting wippersnapper FS code rn and want to ensure I'm not missing anything
onyx hinge
look at filesystem_init in supervisor/shared/filesystem.h
prime flower
@onyx hinge filesystem.h doesn't exist on main?
onyx hinge
well this is a chicken-and-egg problem: The voltage bus called PER_VCC_3V3 is off unless an I2C expander is pulling a particular output LOW. PER_VCC_3V3 powers the I2C pull-up resistors. So, CircuitPython can't create the I2C bus to turn on the pull up
filesystem.c, oops
ornate breach
It can be finicky, and requires a CSV of possible pins for a particular microcon...
Ah.. I don’t have a way to make fritzing images :/
prime flower
@onyx hinge Perfect, this is exactly what I was looking for. Thank you!
idle owl
Ah.. I don’t have a way to make fritzing images :/
So... We also have a repo up for the software we use for making the Fritzing objects. But it takes finicky to a whole new level.
Do you have an EagleCAD .brd file for the intended board?
And a way to edit SVGs
prime flower
@onyx hinge (maybe @tulip sleet ) do you know why it creates metadata_never_index?
idle owl
<@!426200547024961539> (maybe <@!329766224093249548> ) do you know why it creat...
It's a MacOS thing.
onyx hinge
prime flower
👍
slender iron
@onyx hinge I'm at my desk now if you want to sync on this esp issue
onyx hinge
@slender iron I'll have lunch soon .. how's ....:30 (in about 40 minutes) for you?
slender iron
works for me. I'll get through my email
ornate breach
Do you have an EagleCAD .brd file for the intended board?
Yeah, I use eagle
slender iron
remind me to exercise after 🙂
ornate breach
And also yes on the svg part as well
onyx hinge
If an I2C device address is written as 0'b 0100 000x do I want 0b100_000 or 0b100_0000 in Python (0x20 or 0x40)?
neither actually seems to work, mind you
slender iron
I just do a scan...
onyx hinge
scan hangs
so there's something else wrong
or else it's taking a very long time
what's "(NC)" mean on a schematic? is that like non-populated?
idle owl
And also yes on the svg part as well
Ok, there's a guide on creating a Fritzing object using our eagle2fritzing software, but it's probably woefully out of date in some places. https://learn.adafruit.com/make-beautiful-fritzing-parts-with-eagle2fritzing-brd2svg/overview
idle owl
what's "(NC)" mean on a schematic? is that like non-populated?
Not connected.
Or no connection.
In sensor datasheets anyway, that's what it means - don't connect anything to this.
I might be wrong about what it means on schematics though.
slender iron
I think you are right kattni
onyx hinge
OK, it scans now
slender iron
you should be able to find a spot on the pcb labeled that doesn't have the resistor
onyx hinge
OK so verified that this board has NO i2c pull-up resistors, not even the ones that it appeared would be powered when the PER_VCC_3V3 rail was powered.
idle owl
My script is not working like it should be.
idle owl
And I don't understand why.
It's supposed to exclude certain pins, but it is no longer doing that.
idle owl
This is not excluding the pins in exclude. 😕
slender iron
I'd not use the comprehension so I could add prints
idle owl
Limor wrote that part a while back. I'll try to tease it apart.
ember iris
This is not excluding the pins in `exclude`. 😕
what is the type of p? And what does dir(board) return?--so what is the type of an element of dir(board)
idle owl
what is the type of p? And what does dir(board) return?--so what is the type of ...
dir(board) returns a list of pins.
p should be a pin type. I think.
ember iris
are they strings or of a Pin class?
idle owl
not type, pin name.
Hold on.
['__class__', 'A0', 'A1', 'A2', 'A3', 'A4', 'A5', 'APA102_MOSI', 'APA102_SCK', 'BLUE_LED', 'D0', 'D1', 'D10', 'D11', 'D12', 'D13', 'D2', 'D5', 'D7', 'D9', 'DOTSTAR_CLOCK', 'DOTSTAR_DATA', 'I2C', 'L', 'LED', 'MISO', 'MOSI', 'RX', 'SCK', 'SCL', 'SDA', 'SPI', 'SWITCH', 'TX', 'UART']
>>>
Does this, pins = [pin for pin in [getattr(board, p) for p in dir(board) if str(p) not in exclude] if isinstance(pin, Pin)] have any impact?
jaunty juniper
hmmm it works though, check you are running the right file or something ?
what is the issue exactly ? could you be getting pins you don't want because they have alias names that are not in the exclude list ?
ember iris
Well, ... time to drop the list comprehension then. For whatever reason p in dir(boarrd) is not able to compare against the exclude list.
pins = []
for p in dir(board):
print('Checking ', p)
if p not in exclude:
p_attr = getattr(board, p)
if isinstance(p_attr, Pin):
print(p, ' is a pin and')
pins.append(p_attr)
else:
print('Excluding pin', p)
I don't know if the pins are appending the correct thing there--but the prints should be clarifying
idle owl
what is the issue exactly ? could you be getting pins you don't want because the...
Yes, it's returning pins that are in the exclude list.
jaunty juniper
for example if I put "A1" in the exclude list it still gets returned because it is also "D1"
ember iris
Yes, it's returning pins that are in the exclude list.
wait does this mean pin A0 could also be SWITCH
jaunty juniper
in which case you could make the exclude list a list of pins (with getattr to get the base name not the alias) instead of a list of names
idle owl
Ohhh.
I get it.
It was working. But we added new names for the APA pins
so they showed up
This makes a lot more sense.
thank you.
onyx hinge
@slender iron OK I'm done with 🌮
ember iris
Huh, I wouldn't have thought of that, and would have gotten caught in that pitfall soon. Thanks for highlighting it!
slender iron
kk
idle owl
in which case you could make the exclude list a list of pins (with getattr to ge...
Not clear on how I would do this.
tidal kiln
is CP's behavior for getting a NACK on a i2c write to throw an OSError? (it appears to be, but want to confirm)
jaunty juniper
something like: exclude = [getattr(board, p) for p in ["NEOPIXEL", "APA102_MOSI", "APA102_SCK", "L", "SWITCH", "BUTTON", "A1"] if p in dir(board)]
idle owl
Ahh ok
jaunty juniper
and then pins = [pin for pin in [getattr(board, p) for p in dir(board)] if isinstance(pin, Pin) and pin not in exclude]
oh I wanted to use a set() for uniqueness, but Pin is not hashable 😦
idle owl
😕 Yeah, there's some tweaky things that made this annoying to write.
jaunty juniper
makes you jump through hoops
tidal kiln
idle owl
@lone axle Ugh, ok, so PyCharm stopped responding, so I force quit it, and when I restarted it, the project directory structure doesn't show up on the left anymore. And I can't find where to get it to come back. Any ideas?
lone axle
<@!382939733107408897> Ugh, ok, so PyCharm stopped responding, so I force quit i...
If you see this thing along the side still you can click it to hide/show the project pane
alt+1 is a keyboard shortcut for hide/show that pane as well.
idle owl
If you see this thing along the side still you can click it to hide/show the pro...
It only says "pull requests" down the side 😕
And I'm on Mac, even this windows keyboard doesn't seem to have alt anymore.
lone axle
if you don't see that along the side anymore check View -> Tool Windows and try to find 'Project' inside of there
Clicking on that in the menu should bring it back I think. Mine also shows the shortcut keys off to the ride side in that menu, maybe it'll list what you can use instead of alt.
idle owl
OK thanks I'll try that.
idle owl
if you don't see that along the side anymore check View -> Tool Windows and try ...
It wasn't in the list. I tried restarting again, and now it shows up. 😖
Anyway thanks, now I know where to find the shortcuts for those things.
Also shows up in the list again.
So weird.
It's CMD+1, which I tried, and it didn't work.
lone axle
yep, you're welcome. I occasionally get similar issues where stuff is weird when I open up the app and restarting tends to get it back to normal.
idle owl
Huh. Ok. Good to know.
manic glacierBOT
spare prawn
hii sorry for the dumb question. i (stupiidly) managed to get into a reset loop (microcontroller.reset()) in code.py. Is there any way i can break out of it without reflashing the entire device (nrfExpressbboard)
jaunty juniper
hii sorry for the dumb question. i (stupiidly) managed to get into a reset loop...
press reset, then when the status LED blinks yellow, press reset to boot in safe mode, which doesn't run user code
spare prawn
thx!
blissful pollen
hii sorry for the dumb question. i (stupiidly) managed to get into a reset loop...
I think we have all done that before!
solar whale
I think we have all done that before!
Sure have -- but it could be much worse -- there was a spacecraft that had a command to turn it off -- someone executed it ... https://heasarc.gsfc.nasa.gov/docs/heasarc/missions/phobos1.html
idle owl
Sure have -- but it could be much worse -- there was a spacecraft that had a com...
Wow. I'm keeping that one in my pocket to help folks feel better about mistakes!
idle owl
Is there a phrase or word that means "not broken out" with regard to pins?
e.g. "This is not an exhaustive list of not broken out pins."
tulip sleet
Is there a phrase or word that means "not broken out" with regard to pins?
unconnected ?
idle owl
Well they're connected to something, aren't they?
I mean pins you can't use on a microcontroller board because they aren't made available to you with a pin or through-hole.
Trying to comment this audio pin script I wrote in case someone wants to add more excluded pins later.
Other than me.
The pins I'm referring to are connected, they're simply not available to you to connect to directly (without a lot of effort and tiny soldering).
like NEOPIXEL and DOTSTAR_CLOCK/DATA.
etc.
tidal kiln
i think "not broken out" is fine
tidal kiln
"unexposed"? "unavailable"?
tidal kiln
"unbrokenout"?
idle owl
I like unexposed. I think it works.
timber mango
My code which is running on an RP2040 will consistently crash into the hardfault handler when run on the nRF. I've tried removing bits of the code to see which part is causing it, but it seems to just happen at a different spot instead. Is there anything i can do to better troubleshoot crashes to the hardfault handler?
idle owl
My code which is running on an RP2040 will consistently crash into the hardfault...
Bear in mind, you shouldn't be crashing into the hardfault handler. So consider filing an issue on the CircuitPython repo with your code and version info. But, someone else may have better suggestions for troubleshooting it more easily.
timber mango
Yeah, the hardfault handler does mention filing an issue 🙂 But rather then posting a big blob of my horrible code, i would like to try narrow down the problem if all possible, and save everyones time
idle owl
Yeah, the hardfault handler does mention filing an issue 🙂 But rather then post...
I know that feeling entirely. You're doing what I would do, but there may be a better way to do it that I'm unaware of.
ember iris
Oh that's a messy knot of a bug
ember iris
@onyx hinge can I ask a question regarding that bug? I think I hit that bug with my metro esp32-s2 as well, but once I hit a crash that knocked it into safemode, I haven't been able to trigger it (more than once) again. Could it actually have been related, or was I messing something up and just thinking it was related? It would only reset when the memory needed to be freed up
I have no idea why on earth safemode would have kicked it out of the issue
onyx hinge
It's possible
A crash should usually make you enter safe mode but something about the esp32s2 means this bug was almost always making us restart normally, not go to safe mode
I can't predict the future but Scott may talk about it on his next stream
ember iris
Huh. Well what an awesome wealth of stuff I get to learn about going forward
idle owl
@onyx hinge So in place of indicating which pins work for audio, because some boards have a TON, I'm including a script you can run to get a list for yourself. Like I did on the I2C/SPI/UART pages in the Essentials guide, and the I2C template. (The others aren't templatised yet.)
onyx hinge
@idle owl neat, is it working yet?
cunning crypt
<@!426200547024961539> So in place of indicating which pins work for audio, beca...
That UART script came in handy a bit ago!
idle owl
<@330227457296957440> neat, is it working yet?
It always was, turns out it was duplicate pin names messing with me. Neradoc helped me resolve that.
idle owl
@tulip sleet Were you following up with others about this issue? https://forums.adafruit.com/viewtopic.php?f=60&t=182595
@ember iris This answer is excellent! Thank you for following up with this! https://forums.adafruit.com/viewtopic.php?f=60&t=182579
tulip sleet
<@!329766224093249548> Were you following up with others about this issue? https...
That is yet another one. I'll add it to the internal email thread. Thanks.
ember iris
<@!459119350851567626> This answer is excellent! Thank you for following up with...
sweet! I need to follow up, I'm working on the open issue of scd4x right now, I should be able to commit code for that then write a reply to that post. Thank you for letting me know I was on track! It helps a lot!
idle owl
That is yet another one. I'll add it to the internal email thread. Thanks.
May want to consider letting Dave Astels know that we should be notified of those.
tulip sleet
I thought he was cc'd on the email via a list, but I will check.
thorny jay
I miss pair programming
Happy I could bring this bug report to you. ;-)
idle owl
I thought he was cc'd on the email via a list, but I will check.
Oh, I didn't know. He might be.
idle owl
@tidal kiln I saw you mentioning this on the forums go by - FYI it's published: https://learn.adafruit.com/adafruit-scd-40-and-scd-41
tidal kiln
coolio. thanks!
slender iron
I'll definitely talk about this esp bug tomorrow
slender iron
@timber mango attaching a debugger and getting a backtrace from the hardfault handler is usually my next step
tulip sleet
<@456226577798135808> attaching a debugger and getting a backtrace from the hard...
I think @timber mango has narrowed it down to https://github.com/adafruit/circuitpython/issues/5240
onyx hinge
you had corn to harvest?
slender iron
yup yup
ember iris
I was away longer than I expected because I harvested our corn. 🌽
I'm sorry, as a Nebraskan I have tuned into this portion of the conversation
😅
slender iron
I think I got 6 ears before these
boiled them and cut them off the cob to freeze 'em
ember iris
Prepping and freezing corn is awesome! I need to get some and freeze it myself. I've got a tiny freezer so I've skipped it the last few years, but while it's a chore the taste later on is worth it
slender iron
@onyx hinge just pushed the code I had to https://github.com/tannewt/circuitpython/tree/fix_gc_network but need to think it over a bit more while I play some games 🙂
onyx hinge
@ember iris hey, I'm in Nebraksa (Lincoln).
ember iris
<@!459119350851567626> hey, I'm in Nebraksa (Lincoln).
Well hot dog! That was where I went to college, and where my partner is from. I'm from omaha myself, and pretending to fit in while in KC. Excluding the past year and a half, I've been there once ever month or two if not more
onyx hinge
@ember iris if the circumstance arises I'd be happy to have a coffee or other tasty beverage with you. Likewise, we would usually get down to KC from time to time but it hasn't been happening lately.
ember iris
<@!459119350851567626> if the circumstance arises I'd be happy to have a coffee ...
We're due a trip to lincoln shortly--maybe late October, I'd love to meet up! I'll let you know next time we're on our way up! It's such a nice city, minus O street between 11ths and 17th, but that's to be expected
onyx hinge
Sure, we'll see where things are at in October vis a vis doing in person things. Fingers crossed.
ember iris
Sure, we'll see where things are at in October vis a vis doing in person things....
Fingers crossed! But if not, we're up in lincoln pretty frequently, so there's no worry that 'suddenly' things won't line up
onyx hinge
I have an awful idea. what if I create a keypad with 2 keys (A and B) with the keypad module, then use python code to "decode" them as quadrature. Is that fast enough for the rates that people manually spin encoder dials, or is it too slow?
since I'd get an event anytime either A or B changed, ...
tulip sleet
I have an awful idea. what if I create a keypad with 2 keys (A and B) with the k...
It might or might not keep up. You may need to reduce the default interval. I have a feeling it may not keep up, considering that I think we see that already with the more minimal interrupt handling in IncrementalEncoder.
onyx hinge
just spotted https://realpython.com/podcasts/rpp/75/ on my feeds
slender iron
@tulip sleet did you start fixing the countio issue?
tulip sleet
<@!329766224093249548> did you start fixing the countio issue?
I only just looked at the code, and put those comments in the issue to record what I saw. I thought you might have a better handle on why the GPIOTE stuff was the way it was (and whether it was wrong). So go ahead if you want; I don't think I will have a lot of time to do further fixes today.
onyx hinge
github is trying to get me to install this .. I am pretty sure we don't want it.
slender iron
I'm not a huge fan of stale bots
though I could see it being useful for specific issues where we ask for a recheck and then give the reporter a week or two reply before auto-closing
timber mango
@onyx hinge Wasn't actually sure if it was intended, but at least it was not consistent across ports - sorry for not being more clear
solar whale
FYI -- I have implement a temporary workaround where I just check if the first two bytes in the image are 0xff. If so skip the first byte... crude but effective. Now I'll start looking for root case in the core, but I wanted to have a working version on the saola.
onyx hinge
@solar whale basically, if it has one byte that should have come from the previous frame .. but it's supposed to clear out any pending bytes when the frame-starting signal comes. did you file an issue about it yet?
i won't look at it for 2 weeks or more so I'm happy you're peering into the core and trying to find something out
slender iron
I'm looking at 5240 now
onyx hinge
I'm not really here, I'm on vacation 🙂
solar whale
@onyx hinge Enjoy your vacation!!! I'm happy to poke around. We'll revisit this when you get back.
wild niche
what is the current state of multiprocessing in CircuitPython?
tulip sleet
what is the current state of multiprocessing in CircuitPython?
Threads are not enabled. We do not support the second core on the RP2040. There is a continuing discussion about concurrency and event loops here: https://github.com/adafruit/circuitpython/issues/4542 that follows on from an older issue that got somewhat derailed: https://github.com/adafruit/circuitpython/issues/1380
wild niche
@tulip sleet thanks, I will check the links out. have a great weekend
manic glacierBOT
timber mango
I am able to get REPL on nRF to freeze, needing a hard reset to get it working, by pasting a bunch from Mu in paste mode. Doing the same on the RP2040 will not freeze.
It doesn't happen at the same spot, but it will happen eventually. Tried it by pasting ~3300 bytes over 5 lines, but it seems to happen with any amount of lines..
Should i make an issue for this? Or just stop pasting as much? 🙂
jaunty juniper
I have seen that happen once today, while testing if https://github.com/adafruit/circuitpython/issues/3152 was still there (it is)
timber mango
Ah, already an issue there
jaunty juniper
it's not the same issue, this one is not about a freeze
manic glacierBOT
timber mango
it's not the same issue, this one is not about a freeze
True
manic glacierBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
solar whale
FYI -- after updating to the current tip of main, I had followed my usual pull/update process git pull then git submodule sync --recursive git submodule update --init git submodule foreach --recursive 'git fetch --tags' I then also had to delete and re-install/update esp-idf since the folder was out of sync. I tried to build an esp32s2 (saola_wrover) and it failed as follows ```/home/jerryneedell/projects/circuitpython/ports/esp32s2/boards/espressif_saola_1_wrover/sdkconfig...
CMake Error at esp-idf/tools/cmake/component.cmake:305 (message):
Include directory
'/home/jerryneedell/projects/circuitpython/ports/esp32s2/esp-idf/components/mbedtls/mbedtls/include'
is not a directory.
Call Stack (most recent call first):
esp-idf/tools/cmake/component.cmake:477 (__component_add_include_dirs)
esp-idf/components/mbedtls/CMakeLists.txt:4 (idf_component_register)
-- Configuring incomplete, errors occurred!
See also "/home/jerryneedell/projects/circuitpython/ports/esp32s2/build-espressif_saola_1_wrover/esp-idf/CMakeFiles/CMakeOutput.log".
make: *** [Makefile:253: build-espressif_saola_1_wrover/esp-idf/config/sdkconfig.h] Error 1
make: *** Deleting file 'build-espressif_saola_1_wrover/esp-idf/config/sdkconfig.h'
ws ``` I then just recloned the entire circuitpython repository and re-updated everything and the build succeeded, I just wanted to pass along a warning to anyone updating...
manic glacierBOT
manic glacierBOT
manic glacierBOT
atomic summit
Hey folks - Can I please get some eyes on this pinout (for sanity reasons) to confirm I have made the Dx/Ax numbers compatible with the FeatherM4 ? After making my FeatherS2 pinout slightly different (making life hard for folks), I don't want to make the same mistake with my FeatherS2 Neo. Thanks!
onyx hinge
@atomic summit is there psram?
atomic summit
<@!195765176035835904> is there psram?
Yes, 4MB Flash, 2MB PSRAM in the ESP32-S2 chip.
onyx hinge
@atomic summit looks like the pin you have labeled IO45/D2 is D4 on the Feather M4 express
atomic summit
<@!195765176035835904> looks like the pin you have labeled IO45/D2 is D4 on the ...
That's exactly why I need "fresh eyes on this"... totally missed that, thanks!
onyx hinge
You're placing D0 where the Feather Specification states:
AREF - this is the Analog Reference if there is one for the chip. If there is not, keep this pin not-connected. We recommend Wings don't require this pin as not all chips have external ARef's.
If you're doing this knowingly, then do away. Note that Feather STM32F405 puts the BOOT pin where Feather M4 put "D4". That spot is
'Free' Pin - This is the pin to the right of TX. You can use it for an extra GPIO or if you have some onboard module that has a useful breakout. Sometimes we tie it to ground. Do whatever you like! FeatherWings should not require or use this pin unless there's some really good reason.
apparently nobody checked this pin list when they made the feather pico's pins go A0 A1 A2 A3 D24 D25 (i.e., they should have made it D18/D19 to match Feather M4), probably late to fix now.
Adafruit would rather your silk show the D-numbers but again that's something that you may be doing in this way knowingly
atomic summit
IO0 on that AREF pin is required. I do it on the FeatherS2 and I've never heard of any issues with having it there.
onyx hinge
(I mean, instead of the IO numbers. SCK/SDL etc get labeled with their function)
That's all I got in the way of commentary. I hope some of it was helpful!
atomic summit
D numbers is not a thing for ESP32s. IO number is. That's a deliberate thing by me. But having the right D numbers in the CPY firmware for compatibility is what I want to ensure so CPY code works.
I can also add them in Arduino as well in the board definition.
atomic summit
That's all I got in the way of commentary. I hope some of it was helpful!
Yes it was, thanks heaps for looking over it!
D2/D4 on IO0/45 positions was back to front... nice catch
onyx hinge
ah, a swap
atomic summit
Yup - updated image..
spiral elk
@atomic summit I'd like to say I wish my hands were half as steady as yours manually placing 0402s and 0603s...
atomic summit
<@!195765176035835904> I'd like to say I wish my hands were half as steady as yo...
Just takes practice.. LOTS and LOTS of practice 🙂
spiral elk
Or maybe I just need to lay off the coffee
Is that an LED array in the extra space left over from the flash/psram being internal?
atomic summit
Yup, it's a 5x5 RGB Matrix onboard - with user controlled dedicated LDO for power.
spiral elk
I hadn't heard what you intended to do with it, just that you had plans for the space.
atomic summit
Initially my plans was to make a "cheaper FeatherS2" as the silicon shortage is making PSRAM & Flash super expensive and really hard to source... but adding 25x 1515 RGB LEDs ended up costing as much as the Flash & PSRAM - hahahaha
spiral elk
Well, at least you only have availability of the FN4R2 to worry about, but its all eggs in one basket.
atomic summit
And the FN4R2 is in short supply - been in a queue for 6 weeks for more reels, no ETA yet 😦 Lucky I also use it on my TinyS2 so have some stock here.
spiral elk
Lolin is probably buying up a ton of them.
onyx hinge
when re-flashed with upstream's sensor demo the screen is correct. Docs say it's a "Display interface: 16-bit, 20 MHz, 8080 parallel communication" so I suspect that talking to it is just a matter of extending the parallel-bus display to 16 bits
though the I/O pins are nonsequential, that's a wrinkle
manic glacierBOT
onyx hinge
Yup - updated image..
I had a look again at the A / D / busio names compared to metro m4 and didn't spot any other items to mention
manic glacierBOT
manic glacierBOT
mortal mica
Got to listen to one of Scott's deep dives while I was falling asleep last night and woke up to an important discussion between Limor and Scott regarding the supported qspi flash chips for the RP2040. Is there a list of supported/known working chips so that I dont get any surprise. Digikey only had the DTR version of the W25Q128JV. Will it work? Does it need a different config? I tried looking it in the docs or in the code but couldn't find much for an answer.
idle owl
<@&356864093652516868> CircuitPython Weekly meeting is in about 1.5 hours. Please remember, if you wish to participate, or want to have your notes read off during the meeting, you'll need to update your Hug Reports and Status Updates in the notes doc. Thanks! https://docs.google.com/document/d/1bO1IU2lth51Bl0pKTlVy9IYak9HuB7GFfyaRQhFISxs/edit
slender iron
@mortal mica I think all new winbond should be ok. worse case you need to add new config for it
generally you want something that can do quad spi
mortal mica
Thanks. I have parts coming from all over. Might be a while before I know for sure but it sounds like it's going to be ok... Where would a new config need to be added? ( So that I can compare with existing ones so that I know what I would need to get into)
slender iron
thorny jay
<@&356864093652516868> CircuitPython Weekly meeting is in about 1.5 hours. Pleas...
Skipping one more time, testing aluminium tube transportation in pandemic time. 😭
pearl notch
jaunty juniper
idle owl
@lone axle Thanks for keeping up with cookiecutter!
lone axle
same here with organizing pain, I have a good bit of it that still needs to be done as well
idle owl
This is kind of what I was expecting as a response from most folks 😄
gilded cradle
So I have most boards grouped into Feathers, Raspberry Pi Boards, Metro/Grand Central Style, and CPX/Misc
idle owl
So I have most boards grouped into Feathers, Raspberry Pi Boards, Metro/Grand Ce...
I was debating between this or by MCU.
gilded cradle
Also have a breakouts category for add-on boards.
idle owl
I have a lot of everything, so any choice will have enough to fill a bin.
gilded cradle
I think form factor sorting works quite well.
idle owl
Fair enough. I might give that a try.
gilded cradle
The only time I have a hard time finding a board is when I forget to put it back in the right place.
idle owl
Breakouts, I'm trying out doing them by concept, e.g. air quality/gas, temp/humidity, etc. So all of the air quality/gas sensors go in one container, and so on.
gilded cradle
That's how I do it
idle owl
Might need to combine a couple because I don't know if I have enough of every type to fill an entire container, but motion/IMU/accel/gyro and temp/humidity will both definitely have enough 😄
OK, thanks for the suggestions. I think I'll give this a try.
gilded cradle
Yeah, voltage regulators and level shifters are good categories too
oh and FRAM chips
idle owl
My previous breakout organisation was alphabetical by part name. Turns out there's a lot of maybe four letters of the alphabet, and hardly any of any other. It works as a concept, but I think by feature is a better way to go. I'll find out eventually!
gilded cradle
Yeah, easy enough to locate a small pile and look through them
ember iris
Thanks all!
solar whale
Have a good week all 👋
slender iron
@inland tusk https://youtu.be/LPPoc_0_Haw
Here is the notes document for Tuesday's CircuitPython Weekly meeting. It is at the normal time of 11am Pacific / 2pm Eastern here on Discord (just 24 hours later). 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/1Dag1KnCTNh50q5fc_ghaMdmUjNiBp5KsdF4FbKKnIEk/edit?usp=sharing
turbid radish
?serverinfo
digital shoreBOT
adafruit
mortal mica
https://github.com/adafruit/nvm.toml
Thanks. The list is a whole lot longer than I expected.
slender iron
🙂
ember iris
*frantically adjusts pull request* 😄
idle owl
😄 No worries!
ornate breach
frantic programmer noises
idle owl
That smacks of a terrible subtitle provided for some video.
ornate breach
It’s true
manic glacierBOT
manic glacierBOT
atomic summit
I know we are super close to a 7.0 release and 6.3.0 is stable and not being worked on anymore, but is there a way to get my FeatherS2 Neo added to 6.3.0? If I do a PR against the 6.3.0 tag?
orchid basinBOT
manic glacierBOT
tulip sleet
I know we are super close to a 7.0 release and 6.3.0 is stable and not being wor...
We haven't done this even for Adafruit's own boards. 7.0.0 is really better in pretty much all ways. Is it just that people will refuse to use a "beta"?
manic glacierBOT
idle owl
I know we are super close to a 7.0 release and 6.3.0 is stable and not being wor...
PRing to the tag doesn't automatically add it to the version. It would involve doing another release, which isn't planned because 7 is so close to stable at this point. As well, 7 includes many bug fixes and important features. We're supporting 6 at this point because it's the last stable release, but we don't have any intention of backporting anything. As Dan mentioned, we haven't even included new Adafruit boards.
slender iron
my goal is to do a 7.0.0 release candidate this week and if nothing major comes up then we can make it stable next week
idle owl
I thought as much, but didn't want to speak to that without verification from you.
stuck elbow
phew, I got the keypad changes just in time then, I have been putting them off for months
ember iris
So I'm reviewing a pull request, and because one of the other sensors it uses as an example has changed, the example has become out of date. But it's not because the pull request is wrong. So should I request a change or raise an issue, then fix the sensor in the example with a separate pull request?
tidal kiln
@ember iris can you link to pr
ember iris
Here's the pull request: https://github.com/adafruit/Adafruit_CircuitPython_SGP40/pull/9
and this is the senor that's changed: https://github.com/adafruit/Adafruit_CircuitPython_BME280
The import for the bme280 needs to be changed from:
import adafruit_bme280
to
from adafruit_bme280 import basic as adafruit_bme280
tidal kiln
oh. in the readme?
idle owl
So I'm reviewing a pull request, and because one of the other sensors it uses as...
I think it depends on your feelings on the person who filed the PR - which is to say, do you feel they're up for making the changes? Or are they new and struggled with what they submitted in the first place and you don't want to put more on them.
tidal kiln
^^ yah. agree.
idle owl
This looks like a PR in which the original author could make the changes comfortably.
Basically, if a new person contributes, and it's something little that you think should be changed, but what they submitted was already a huge deal to them, you can approve it and make the change afterwards.
Or something big you think should be changed. Either way.
ember iris
I think it depends on your feelings on the person who filed the PR - which is to...
I hadn't thought of that, that's a good point
idle owl
You can also, always ask!
I've asked folks when I'm unsure in the PR "Do you feel comfortable making this change?" and sometimes they say no. So I'll do it either in that PR or at a later time with a new PR.
Other times, they're new, and they're totally into it, and want to make the change. And they'll stick with you as you ask for all sorts of stuff, or fight with the CI with them, and so on.
But you don't want to frighten off new folks with an onslaught of change requests without feeling it out first.
ember iris
Sweet! I just completely didn't think of it from that point of view. I was focused on the, "should a PR's scope be small or larger" perspective
idle owl
That's a valid perspective as well though!
Adding the change you're suggesting doesn't bloat the scope too much. But if it's something unrelated, or much bigger, you could consider a separate PR in that case regardless of how the original author is feeling about it.
Or, again, ask the author if they are willing to submit another PR with the scope-enbiggening change to avoid one PR being too big.
ember iris
Ok.
In a similar vein, I have one more question regarding the sgp40--it reads a raw value, then you pass that value to the algorithm which spits out a voc index. Storing that raw value adds a small memory footprint (the size of an int plus the class overhead for it). I'm probably like one of 5 people who would be interested in that value. Is it worth adjusting the class to save the raw value, or should I just make my own fork that has that adjustment?
tidal kiln
can voc index call raw? why pass it?
idle owl
How about figuring out the actual memory footprint it creates so we have numbers to work with and then we can decide based on actual data.
ember iris
can voc index call raw? why pass it?
Boy we're stretching my knowledge of how these variables and functions work: To the best of my knowledge in this case raw is a property that calls the i2c bus with a stored command_buffer bytes array, and it just returns the value in read_value[0]. Measure_raw adjusts the command buffer for the temp and humidity, and returns the raw property, which is evaluated on the return. measure_index calls measure_raw and stores the returned read_value locally in that function. I'm weak when it comes to understanding name space scopes, but I think that means that I can't access that the raw value from the sensor once that measure_index function returns.
ember iris
How about figuring out the actual memory footprint it creates so we have numbers...
Good idea!
tidal kiln
if you want raw, use measure_raw?
are you wanting both voc index and raw to be returned by a single function?
ember iris
are you wanting both voc index and raw to be returned by a single function?
I'm wanting to be able to access both, but voc index should be the only thing that measure_index returns. But I'm looking at the code and realizing I can just write a function to mimic measure index in my own program, so all of the above is moot
namespaces confuse me 😛
tidal kiln
it looks like it's basically:
vox_index = some_voc_algorithm(raw)
?
it terms of just what is going on internally
ember iris
Yeah, assuming some_voc_algorithm is a class attached to the sgp40 class so it's internal states are preserved
tidal kiln
totally a different approach, but looks like maybe these could be kept separate? that is, don't try and make it a part of the SGP40 class.
you could import that algorithm class separately in your user code
ember iris
That would work, but it would change the behavior pretty dramatically from the c implementation, and the sgp40 documentation. Functionally it's identical, but from an api standpoint it would deviate
tidal kiln
import board
import adafruit_sgp40
import adafruit_sgp40.voc_algorithm
sgp40 = adafruit_sgp40.SGP40(board.I2C())
voc_algo = adafruit_sgp40.voc_algorithm.VOCAlgorithm()
raw = sgp40.raw
vox_index = voc_algo.vocalgorithm_process(raw)
yah, it'd be different
ember iris
And now I've hit another issue. Can I ask for a second set of eyes on this?
This is a time based algorithm:
But I don't see where this file can know the time that has elapsed. Am I just missing it?
tidal kiln
sry. i'm probably not helping here. i don't really know anything about that algorithm. looks like it's more than just math(raw).
ember iris
sry. i'm probably not helping here. i don't really know anything about that algo...
No worries, I was going to redo this algorithm with floating point instead of fixed point math, but the pull request was made with the fixed point implementation solving the open issue.
The only reason I caught the time issue is because I was trying to understand the distributions and estimators and there's a 2 hour switch. I'll dig in more and see if I'm missing something or if this breaks if the sgp40 is sampled at a rate other than 1 once a second
mortal mica
Been looking at sleeping on the nrf52 and was wondering why it was decided to use GPIOTE instead of PIN Sense for the wake up mechanism (especially for deep sleep). Considering the limited number of GPIOTE channels and the "unlimited" number of PIN Sense "channels" and the fact that a transition on a PIN when in SYSTEM_OFF mode will wake up the chip (from datasheet: POWER Peripheral uses the DETECT signal to exit from System OFF mode). I understand if I want to wake up on a time basis. Perhaps we need a "hibernate" mode where external input is needed to wake up. (I am deciding if this something worth looking into and contributing. Would it be taken in and where/how?)
onyx hinge
@mortal mica it's entirely possible we picked a less than ideal way to implement the sleep mode on that MCU, I don't know enough to say .. if there's a different implementation that can implement the python API but it's better in some way like lower power or additional flexibility I think we'd love to see a pull request
mortal mica
Great. I wonder about additional flexibility. Thinking more like different flexibility... Is the python API what's in the docs? or something else... For example, I don't think there is "Hibernate" anywhere already defined.
As for power, its lower but going from 180uA down to 60uA is something but I am more worried about when it's running (see http://pykey.jpconstantineau.com/docs/testing_hardware/sleep/). However being able to awake from more pins is the thought here. (would link to a keyboard matrix, energize it, set up the sense pins and go hibernate..)
slender iron
I think I found 1k bytes for samd21 builds at the expense of slower startup
(by computing the qstr attr table on startup)
props to tyomitch for the idea
tests it
ember iris
So does that shunt the qstr attr table into a different type of memory?
atomic summit
Morning @tulip sleet & @idle owl - Yup, that all makes sense and is totally fine. I just wan't sure if the board would just get picked up by "a build system" or not, hence me asking.
I'm shipping with my own 6.3.0 build as I've been building stock for a little while, and when I started, 7 still had that code.py reload issue on ESP32S2, but folks can upgrade them to 7 if/when they like.
idle owl
Anybody up on I2S audio playback? More specifically, using the UDA1334A (https://www.adafruit.com/product/3678) ? I don't have 3.5mm plug enabled speakers, and I'm trying to figure out if I can wire this https://www.adafruit.com/product/4445 to the breadboard pins on it. Says it's meant to be used with an amp, but can be used without one, so I'm unclear. 😕
The only amp breakout I have is the PAM8302, and I don't think they're meant to be used together.
ember iris
I believe you can, it's just going to be somewhat quiet (wait until you know it's quiet before leaning your ear close to it though--learned that one the hard way)
Though I've only ever done that with 8 ohm speakers + a 3.5mm plug.. I think it should work none the less
idle owl
I believe you can, it's just going to be somewhat quiet (wait until you know it'...
Can hook that speaker up to the breadboard pins?
ember iris
yeah--give me like five minutes and I'll have a picture too. Speakers ground should go to the 3.5 ground, and speakers positive should go to one of the other 3.5mm pins, it'll play the left or the right side of the audio accordingly
idle owl
Oh those are the same pins. Obviously. 🤦🏻♀️
I'm thinking, I can't wire it to the 3.5mm, I don't have the hardware.
But you're saying the breakout pins are the same pins.
ember iris
yeah the rout or lout should work I think--they're the right and left channel respectively, so if you wire the red of the speaker to either one of those it should work with the other wire connected to ground AND if anyone has experience to overwrite this let me know. I'm not an audio person
Now I'm magic blue smoke second guessing it..