one that would handle the differences in buttons, sounds, displays, and loading of assets, and give the programmer a consistent api for all the supported boards
#circuitpython-dev
stuck elbow
it could also handle "packaging" of the games and their libraries and assets
and a game selection menu
so you would just copy a directory like you do with /lib right now
but I really don't have the energy to work on something like this right now
lone axle
Agree'd that would be awesome. I don't think I am capable of all the parts to it though. But I'm up for trying to tackle any parts that I can. Even a relatively easy to use solution for buttons would be a great start to get us back to similar capability that was possible before. I will do some experimenting and add thoughts to that issue post once I have some more concrete ideas tried.
<@&356864093652516868> the weekly meeting will occur here on the discord in a bit under 90 minutes from now at 2pm eastern time. The notes document is available here: https://docs.google.com/document/d/1yqZABCU8BjP_njl0_8fed1P2dkDmNqw5o42qjZzLxqc/edit?usp=sharing if you have the chance go ahead and add your hug reports and status updates to it. We look forward to talking with everyone then!
stuck elbow
I'm leaning towards a class that on init inspects the board module and guesses the buttons, then gives you the api and the constants – because then you can do hacks like with the analog joystick or even accelerometer or touchscreen
slender iron
pylint folks. does this error look familiar? https://github.com/adafruit/Adafruit_Blinka_bleio/actions/runs/3959041983/jobs/6781347211
lone axle
pylint folks. does this error look familiar? <https://github.com/adafruit/Adafru...
formatargspec specifically does ring a bell for me. But I don't recall much about the fix beyond maybe having just upgraded to a newer pylint version. Maybe @proven garnet may remember anything more about it from this time #circuitpython-dev message
ember iris
I think you moused over the chatbox icon and that dropped you into the video tiled page
analog bridge
@tulip sleet Before you make a release, #7477 fixes a bug in create_website_pr.yml.
onyx hinge
@idle owl I notice that for me the "turn on camera" button is available, but maybe that's only for 'high roles'...?
(specifically in the voice channel)
idle owl
<@330227457296957440> I notice that for me the "turn on camera" button is availa...
I'm thinking that too.
I think mods are exempt.
tulip sleet
<@329766224093249548> Before you make a release, #7477 fixes a bug in `create_we...
Thanks. We are still a few days away from an rc.0.
analog bridge
text-only
manic glacierBOT
midnight ember
Congrats on 100 PR's to TCFranks! 🎉
proven garnet
`formatargspec` specifically does ring a bell for me. But I don't recall much ab...
Yes! Mismatch between pylint and Python version.
idle owl
@still zephyr Are you text only for your status update as well?
idle owl
No worries!
still zephyr
🙂
onyx hinge
wonders how many PRs the @jepler github account has made over the years
onyx hinge
well also times a large number of years helps 🙂
thorny jay
Sometimes we call the Ada Bot???
proven garnet
Congrats on 100 PR's to TCFranks! 🎉
Not at the meeting today but congrats!
midnight ember
@lone axle made some awesome progress with porting Displayio features to Blinka, that was a really interesting stream.
Hope Tekktrik and Melissa and everyone else dealing with stuff gets well soon. ❤️
idle owl
My GitHub activity, apparently:
9732 Commits
974 Pull requests reviewed
395 Pull requests opened
44 Issues opened
499 issue comments
44 seems low for issues, but who knows.
midnight ember
I think last week Dan said some feature was available in IDF 5.0 but if we're on 4.4 does that mean we can't get those updates until we can port to 5.0?
thorny jay
My C64 to USB experiment.
analog bridge
@tulip sleet give a shot to the IDF v5.0 PR on a non-psram board and see if that fixes stuff
thorny jay
My Argon Light 2.
idle owl
My GitHub activity, apparently:
9732 Commits
974 Pull requests reviewed
395 Pull...
This may be entirely inaccurate. I regenerated the same info, and it's entirely different. 🤷🏻♀️
still zephyr
I have only 3 issues opened 
onyx hinge
My GitHub activity, apparently:
9732 Commits
974 Pull requests reviewed
395 Pull...
where did you get that info? I was busy writing a custom script 
idle owl
where did you get that info? I was busy writing a custom script <a:headdesk:4456...
Hah! I went to a site I found when writing the GitHub profile guide, that has a web instance of their metrics stuff.
midnight ember
Is there are a way to spit out contributions per person like that? That would be cool to see. I think I'm at 2.
talandsia!
midnight ember
Just spritz them using a water bottle with a misting nozzle. In nature they pull mist from the air, it's what they're called air plants.
They grow naturally here in FL.
idle owl
Just spritz them using a water bottle with a misting nozzle. In nature they pull...
I found the soak method in multiple instructions.... Including from the place I bought them. I'll try to keep up with that.
ember iris
Woo good luck with plants! They look lovely!
midnight ember
Beautiful plans!
idle owl
I do have food for them in a spray bottle though!
Was supposed to do that after watering, oops. I'll do it when we're done here.
midnight ember
Air plants when they bloom look gorgeous, similar to bromeliads in structure.
idle owl
I also read somewhere they only bloom once ever?
midnight ember
depends on species, some bloom annually, every couple years.
idle owl
I grabbed the only two with the magenta bits on them, all of the others were green only, so I assume had not reached maturity and therefore blooming time yet.
Oh. Nice ok.
That's pretty neat they grow in FL.
turbid radish
YES< Please
still zephyr
Pro gif too 🙂
turbid radish
For me they don't have to be big, 128x128 for now
idle owl
Is there are a way to spit out contributions per person like that? That would b...
Yes.
@onyx hinge @midnight ember Their web instance is here. It's limited because it's a free web instance, which is to say some of their fancier plugins don't work. But it definitely gives you basics. I don't know why I'm getting different numbers though. https://metrics.lecoq.io/
blissful pollen
I'll take a look at gifs then. Figuring out how to keep them changing (almost a gif.next_frame()) call may be needed.
midnight ember
Scott super hacking bangle over the air. Most impressive.
idle owl
Also it's limited by requests per hour or whatever. So if others are also generating metrics, it might be a bit slow.
still zephyr
gifs could help to update graphical representation of sensors easily
thorny jay
Suddenly the price of bangle.js watch explode and no one knows why...
slender iron
Suddenly the price of bangle.js watch explode and no one knows why...
price from where? from them directly?
idle owl
price from where? from them directly?
I think he's referring to your work making them suddenly expensive.
onyx hinge
<@426200547024961539> <@126221640559362048> Their web instance is here. It's lim...
"Opened 1523 pull requests" according to that website
thorny jay
price from where? from them directly?
I was assuming some will have a suddent interest in the watch...
idle owl
"Opened 1523 pull requests" according to that website
Way ahead of me, evidently.
analog bridge
Go CP9 
onyx hinge
@idle owl my oldest activity is a decade before yours, so per year I think you're ahead
idle owl
<@330227457296957440> my oldest activity is a decade before yours, so per year I...
Hmm. Fair!
onyx hinge
It would be nice if the 9 cycle was shorter than the 8 cycle has been.
jaunty juniper
I'm sure moving to MP 1.19 and IDF 5.0 will be super smooth 😬
lone axle
(i'll add timestamps for in the weeds and the first topic after the fact. forgot it for those)
idle owl
(i'll add timestamps for in the weeds and the first topic after the fact. forgot...
No worries! They get missed from time to time.
jaunty juniper
I was testing watchdog things and noticed that since the watchdog can't be disabled, it is very disruptive during development, even just autoreload
the forum discussion was barely about that, it's more things I noticed when writing test code to check the watchdog use
midnight ember
my internet keeps crashing 😦
gilded cradle
Thanks everyone
analog bridge
Thanks!
still zephyr
Thanks! Have a good week
midnight ember
Thanks for hosting FoamyGuy. See everyone next week.
turbid radish
Thanks all!
idle owl
@lone axle The final timestamp looks like your stamper reset.
ember iris
thanks all! Have a good week!
turbid radish
Thanks!
lone axle
Here is the notes document for next Monday’s CircuitPython Weekly meeting. It is at the normal time of 11am Pacific / 2pm Eastern here on Discord. Everyone is encouraged to attend! Please add your hug reports and status updates even if you’ll be attending the meeting - it’s super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and we’ll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/1kv7EzMdwVLd8ODEcqrHhSPrz_MGrS5Q9qMI6CgXBbXg/edit?usp=sharing
jaunty juniper
@wraith crow does your do dynamic imports and such ? I am wondering is this is a long-lived memory issue, if that hypothesis has not been mentioned before ?
see https://github.com/adafruit/circuitpython/issues/2687
with the use of SPI RAM triggering an edge case or something
just an idea though, but when I saw variables being misplaced I thought of that
wraith crow
Thanks, I'll have to stare at that issue for a while but any leads at this point 🙂
I think I do indeed use similar imports. I'll have to check if there's any correlation with the corrupted variables and the imported ones. I've run this code on half a dozen ports and dozens of boards without running into this so there must be a strange edge case if this is related.
I've got to run for a bit but tonight I'll build the Tinypico bits with and without the SPI RAM and test out the example code from #2687.
jaunty juniper
I don't know if the "long lived" system can be disabled on compile, that could help establish whether there's a correlation ( @slender iron )
slender iron
there isn't a compile time setting for it
tiny hare
Is there any way to contribute to a library/make a request/change some code? Cuz currently, for the circuitpython MCP2515 only 8 MHz and 16 MHz crystal options are available and I've seen some other things that could be improved in that library
Most prefferably I'd just write the code but I don't really know if that's possible
blissful pollen
Is there any way to contribute to a library/make a request/change some code? Cuz...
Hey, all the libraries accept pull requests with changes. They will be tested/reviewed by other community members and then merged in.
There is a guide about contributing to CircuitPython and more information here https://circuitpython.org/contributing
Also in this channel you will find tons of help on getting started and any issues you could come across.
tiny hare
K, thx, how do I make a pull request? I have used git quite a lot recently but I've never contributed to any repository of someone else
blissful pollen
Clone the library you want. Set the clone's origin to the adafruit repo. Then you can create a new branch and do your changes. When you are done you can push back to the adafruit repo.
That guide has a lot better / detailed information on it
tiny hare
Oh, thx great
tidal kiln
there's a lot of nomenclature involved. this might help to understand the overall flow of things.
ideally that would also show making a "branch" somewhere after step 2
the guides discuss branching
jaunty juniper
the key is to work on a branch or it might be painful later down the line
tidal kiln
also - consider opening issue(s) in the library repo first
tiny hare
the key is to work on a branch or it might be painful later down the line
Yep, that I know already
tidal kiln
that helps give the pull request context..."this fixes issue number x"
tiny hare
also - consider opening issue(s) in the library repo first
K, I'll look into that tomorrow
Well, It's not a bug I'm experiencing, it's more of a feature, that I'm missing
tidal kiln
yep. github only has "issues". but its generally ok to open a new issue even it's essentially a feature request.
jaunty juniper
the terms close and resolve are also allowed to auto link an issue from a PR
tiny hare
yep. github only has "issues". but its generally ok to open a new issue even it'...
Ye, I guessed that, just wanted to say...
tidal kiln
and keep each pull request as simple as possible so the code review is easier
jaunty juniper
there's an enhancement tag for a reason 😉
tiny hare
and keep each pull request as simple as possible so the code review is easier
The whole feature added will be a small change
tidal kiln
yep - just for future ref. sometimes people make PR's that add like 5 new features, etc. in a single PR. and the code diff (what needs to be reviewed) ends up being large.
jaunty juniper
my position would be that if you have a PR for a new feature, there's no need to open an issue (whereas bugs might need more of a discussion thread for tests and research)
tidal kiln
like the "some other things" you mention above - defer those to later PRs
jaunty juniper
(or a partial implementation of a larger idea, would need an issue too)
tiny hare
like the "some other things" you mention above - defer those to later PRs
The "some other things" is just an if that checks if the crystal frequency entered is valid but it uses a separate array instead of the .keys func on the crystal dictionary of whatever the function is called to check if the crystal frequency is valid
tiny hare
(or a partial implementation of a larger idea, would need an issue too)
Makes sense
lavish saffron
Hi, the python.org docs for socket.socket describe the args as class socket.socket(family=AF_INET, type=SOCK_STREAM, proto=0, fileno=None) but type is a Python builtin, so how should the type arg be named in a pure python implementation of socket.socket? Thx!
tulip sleet
Hi, the python.org docs for `socket.socket` describe the args as `class socket.s...
It is named type in CPython too, so it's not really an issue: https://docs.python.org/3/library/socket.html?highlight=socket#module-socket. Source code here: https://github.com/python/cpython/blob/5964b1282919b4f82fbe7f4afd28624a4564dc04/Lib/socket.py#L220. As long as the routine doesn't need the builtin name, it just overrides it.
you could argue they (the CPython socket authors) should have named it something else, but they didn't, so we'll just go along with what they did
you can always use builtins.type() even if type has been eclipsed by the arg name
manic glacierBOT
lavish saffron
you can always use `builtins.type()` even if `type` has been eclipsed by the arg...
Thank you!
proven garnet
You may need to suppress the pylint warning where that happens then, just a heads up!
spiral elk
Would there be any violent opposition for me to add a conditional target in the espressif makefile to set some efuses?
(for EFUSE_VDD_SPI_AS_GPIO on C3)
analog bridge
Yes, any firmware loaded in the former dualbank partition will be over written when user_fs gets exhausted.
However, if the filesystem was extended but the extra space was never used then the dualbank partition is still preserved and can be used to boot from.
A new UF2 install always writes ota_0 which is fine if ota_0 is the boot partition but corrupts data if ota_1 is the boot partition and ota_0 is being used as extended storage. I did start work on making TinyUF2 dynamically select and load the current boot partition instead of defaulting to ota_0 but didn't quite complete it.
analog bridge
Would there be any violent opposition for me to add a conditional target in the ...
Is this a sdkconfig setting? If so it can go in the board specific sdkconfig file.
jaunty juniper
the scenario would then be someone using dualbank to update CP over the air, then format the drive into extended partition mode, and then update CP through tinyuf2
by the way I was trying to make a no_ota partition file for 4MB boards with a larger CP partition to fit debug builds but I can't figure out what exactly needs to be done to do that
analog bridge
Checkout PR #6927, it has changes to make a no_ota partition work when CIRCUITPY_DUALBANK=0.
jaunty juniper
ah ! thanks !
analog bridge
Would there be any violent opposition for me to add a conditional target in the ...
We can have a general implementation in the makefile which looks in the board specific config files for what EFUSE bits to set.
spiral elk
We can have a general implementation in the makefile which looks in the board sp...
That's kind of what I was looking for, put what efuses need setting in mpconfigboard.mk, etc.
Though right now there's only the one board that needs that one efuse
tiny hare
what's the preffered way of using git add I allways just do git add . before my commit, is that ok?
I've only changed one file but I'm done now and want to commit it, push it to github and create a pull request
timber mango
Hi
I am emberrassed to make this post but here it goes.
I have used RPis and python for a few years, but never really understood the difference between different python versions, the difference between pip and pip3 or most of the errors I get when I fail to install stuff.
I really want to get OpenCV installed, and I have tried a few times and read instructions on many forums. I have no doubt that OpenCV is great, as well as many of the instructions that I have read, and that the fault lies in me. It is so frustrating knowing that if a knowledgable person would sit at my computer for 2 minutes, everything would work!
So please help me to get OpenCV installed. I understand that to be able to help me, you need to know a few different things. Can we start out by someone telling me what you need to know and how I can get that info?
jaunty juniper
Hi
I am emberrassed to make this post but here it goes.
I have used RPis and pyt...
hi, you want to install OpenCV on a Raspberry Pi ? That would be better asked in #help-with-linux-sbcs, this here channel is for the core development of Circuitpython, so the right people won't be seeing your question here
timber mango
thanks, I will try that
rugged spindle
this might sound like madness but I would like to port pydash to CP - the plan would be to get it all working and then split it up into a library of libraries than can be imported without all the rest of the library - thereby saving large amounts of memory.
however, this means I need to test some base python tools across
some have been no issue but I am having an issue with line 283 of functools.py
def __new__(cls, func, /, *args, **keywords):
which reports:
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/lib/functools.py", line 283
SyntaxError: invalid syntax```
any ideas?this is 7.3.3 on an S2 Mini
rugged spindle
oh? if this isn't possible, I might fork pydash and un-modularise-ise (sic) into different importable helpers that can be adapted
stuck elbow
rugged spindle
ooh yeah - good plan
rugged spindle
right - back to the drawing board. thanks anyway
onyx hinge
this might sound like madness but I would like to port pydash to CP - the plan w...
@rugged spindle https://docs.micropython.org/en/latest/differences/python_38.html Python 3.8's PEP 570 Positional-only arguments is not supported in MP/CP. You could consider def __new__(cls, __func, *args, **keywords): as an alternative, using __ to prevent any "normal" keyword argument from being associated with the positional argument. (and of course changing the use sites within the function itself)
rugged spindle
@onyx hinge I did start doing that but then it just started spiralling into so many that I am rethinking it (as well as my direction in life, etc) 😄
Replicating some hand picked functions will likely end up easier.
onyx hinge
Yeah, there's a lot more complexity to the cpython definition of functools.partial compared to micropython-lib's!
note you can just ask chatgpt to rewrite any code you need to be "more enterprisey" if you need to complicate it later without human intervention
rugged spindle
Hehe nice
lone axle
what's the preffered way of using `git add` I allways just do `git add .` before...
I think the way you do it is fine. You can specify the exact file that you changed if you want but it's not necessary. The main "risk" with doing it your way is if you did change more than 1 thing all of your changes are going to end up getting committed. If that could be a problem for you then it's better to be more specific with your add. But if you are actually only changing 1 thing, then it'll find it and add for you. I do git add --all as well when I am first adding stuff to a repo. I don't know for certain if it's functionally different from your way or not but I think is similar if not the same.
tulip sleet
what's the preffered way of using `git add` I allways just do `git add .` before...
you can also do git commit -a ..., though it does not add new files.
manic glacierBOT
manic glacierBOT
manic glacierBOT
timid bolt
Greetings async fans! I have put together this brainstorm doc of what async APIs might look like. Please comment and discuss. (You can comment directly in the doc.) I'm looking forward to your feedback.
https://docs.google.com/document/d/1v-0MbuARpYzmBExOID4HpNxOTmXXtO4cOYtfreebcWc/edit?usp=sharing
slender iron
I'd definitely add new modules for the APIs instead of expanding the existing API
timid bolt
Yep, that's a valid point. Would you add new modules even if the API changes were non-breaking?
slender iron
probably because small boards likely can't fit the extra code
and we like to keep them enabled/disabled unit at the module level
that way it errors on import
not on use
timid bolt
I see. So it is expected that asyncio is an optional module, and not supported on some boards?
slender iron
@idle owl what's the best way to fix an issue with pylint/cpython incompatibility in the CI? https://github.com/adafruit/Adafruit_Blinka_bleio/actions/runs/3959041983/jobs/6781347211#step:12:95
slender iron
I see. So it is expected that asyncio is an optional module, and not supported o...
Probably. The SAMD21s are basically full
very cool to see the async apis though
we may want to run them by the MP folks too. this may be a chance for us to move to a shared hardware api
timid bolt
I see. How would you structure the new modules. A separate aynsc module for each core module (e.g., digitalio_async, etc) or dump everything in asyncio module, or something else?
slender iron
ya, probably an async version of each module
https://docs.circuitpython.org/en/latest/docs/design_guide.html may be interesting for you to read
jaunty juniper
python seems to use aio for async libraries sometimes ? or is it just aiohttp ?
slender iron
mp has aioble too
tulip sleet
I see. So it is expected that asyncio is an optional module, and not supported o...
asyncio is "just" a library. support for async/await is what is turned off on the smallest builds. There is also _asyncio, which is a C level support for some things in asyncio, but it's optional, and isn't required to use asyncio.
timid bolt
In practice though you need an event loop to do anything useful with async/await, so you end up importing asyncio anyway.
stuck elbow
you can always make your own loop, I did before there was asyncio in cp
with blackjack
tulip sleet
we have had simpler event-loop libraries:
https://github.com/deshipu/meanwhile
https://github.com/bboser/eventio
jaunty juniper
formerly tasko:
https://github.com/kvc0/CircuitPython_async/
timid bolt
Right. So then the trouble is that any async device lib has to late-bind to whatever event loop lib you're using. If you want to support "bring your own event loop" instead of relying on a core event loop impl.
stuck elbow
what kind of binding do you mean? there are well-defined python-level magic functions for all of that
the only thing that needs the low-level code is the select-like call in your loop if you don't want to use polling
and even that could probably be done with events
timid bolt
If I'm implementing aiodigitialio (or should that be digitalaio) for a particular device. The code in this module needs to interact with the event loop, but the event loop impl is not statically know. Previous version of regular Python used to pass around the event loop object. But then they got rid of that, which on one hand makes things simpler.
timid bolt
If I'm implementing aiodigitialio (or should that be digitalaio) for a particula...
Actually maybe they didn't get of the loop parameter. I was reading the documentation wrong.
tulip sleet
I think your document is assuming we can do native (C)async routines. Again, it's not clear to me that's easy or even possible. We are going to talk to the MPy folks soon, we think, and will discuss that with them again.
timid bolt
Nope, there are no native async routines. Async methods are an illusion created at the Python level; they don't exist at the C level.
tulip sleet
but i mean:
class DigitalInOut:
def __await__(self) -> Iterator[None]
async def wait_high(self) -> Awaitable[None]:
while not self.value:
await self
DigitalInOut is native, so how is wait_high() implemented?
timid bolt
My approach to this is to first work out what is the best user-facing API to present to users. Once we agree on that, we can then work out how to implement it. The implementation is the easy part. Understanding users' needs is the hard part. 😀
tulip sleet
i agree, but it needs to be possible to implement.
that was what motivated my passing Events around or returning Events
we can implement some of these libraries in Python, and have a less convenient lower-level interface that is native (that uses Events, say)
timid bolt
wait_high is purely implement in Python. The __await___ method will call C code that will setup the interrupt, then when the interrupt fires there will be more C code that tells the event loop that interrupt happened.
To support the "bring your own event loop" model, __await__ would instead have to be wait(loop=None) method.
stuck elbow
you can't have a method implemented in python on a native C class
timid bolt
The passing around some sort of context object is the right idea. It could work with the MP Event object, but I think the actual event loop would be more fundamental.
tulip sleet
We don't have mixed C/Python native modules. ... ah, so you are not saying digitalio.DigitalInOut, you are talking about digital*a*io.DigitalinOut?
The event loop is invisible in most cases, and I think we can keep it that way. Making the event loop highly visible was a design mistake in old asyncio
and you can see how they have tried to get rid of that
coarse merlin
crikeys. and i'm just sitting here trying to figure out how to convert rpi.GPIO code to circuitpython's digitalio part.
idle owl
<@330227457296957440> what's the best way to fix an issue with pylint/cpython in...
Oof. I'll look into it.
timid bolt
It's a hypothetical DigitalInOut class that supports async. Whether that is the existing class or an new async variant is an orthogonal design decision. You can always manually implement Python code for wait_high in C in a C module.
tulip sleet
@timid bolt I apologize if I am sounding negative. That is not my intention. We are constrained by implementation issues.
The are kind of two aspects of async-ifying the native modules. For some classes, like digitalio.DigitalInOut and countio.Counter, the current routines to get state or do things are not blocking. So right now you can monitor a DIgitalInOut in an async task, and poll it on task switching. That's what's in the current async guide. Adding something like Event support would make that more efficient.
For other things, like socket operations or buisio.I2C.read/write, some audio playing, etc.there are operations that block for a significant length of time. There may be no polling support.
So we could write digitalaio.DigitalInOut right now , In Python, and implement an API that has async support. We cannot write busaio.I2C right now, because we have no non-blocking-with-polling support.
So it would be possible to prototype an API for a lot of this now, and then figure out how to make it more efficient later. We could also experimentally add native non-blocking-with-polling for, say busio.I2C now, either adding to the original module (but we don't have room), or adding a _busaio.I2C that is lower-level non-blocking but non-async, and then implement true async routines on top of that in python
timid bolt
I agree, the visible event loop is annoying. But I think it is necessary to support user supplied event loop implementations, unfortunately.
No not negative all. I really appreciate your feedback and arguments. I have to run some errands right now, but I will happily address your questions when I get back. Thanks.
tulip sleet
thank you too for thinking about this and having a different perspective
slender iron
Oof. I'll look into it.
Thank you!
blissful pollen
Curious if anyone has tried any profiling tools on the CP core? Looking at this mostly for my own learning and curiosity and just wondered if anyone else had looked at it.
slender iron
I haven't. It crossed my mind to run CP in an emulator and just see how many instructions it takes to complete a code.py execution
jaunty juniper
I still dream of using the wokwi pico emulator for profiling memory use and unit tests
and that too
stuck elbow
if we go the route of separate python module with async interfaces to the hardware, I would strongly prefer if there were no name conflicts with the non-async ones, like DigitalInOut. They wouldn't be drop-in replacements, so there is no value in naming them the same, but a huge threat of confusion.
tulip sleet
if we go the route of separate python module with async interfaces to the hardwa...
I agree with this: AsyncDigitalInOut etc or whatever
idle owl
Thank you!
I know what the problem is, and am sorting out the fix. I'll update you once it's sorted.
stuck elbow
I agree with this: `AsyncDigitalInOut` etc or whatever
or even PinChange, since that's what it would effectively be for
I don't think we need async for output
well, not for individual pins, anyways
slender iron
yup, good idea
@timid bolt I'd like to see busio in your async doc too. Maybe some discussion of the internal API too. I'd love to transmit SPI data to a display while computing the next set of pixels.
stuck elbow
<@1065334137243836517> I'd like to see `busio` in your async doc too. Maybe some...
I did some experimenting with that a while ago, and it didn't really speed things up much
the transmission speeds are pretty good compared to the time spent calculating the pixels
I would love to have displays with fewer but bigger pixels
timid bolt
Agreed, there are two cases for async-ification: digitialio vs busio. Some busio-like classes, e.g., UART, have an in_waiting method that you can poll before calling receive which then won't block. This effectively transforms the second case into the first. But I notice I2C and SPI classes don't have this in_waiting method.
Using an async task with Event (at least in it's current form) does not removing the polling, it just pushes it into the C implementation of the event loop. If we are cool with this polling, then I agree there is not much to do. If we could add the "in_waiting" type methods to I2C and SPI we'd be in a good position.
However, the more advanced goal is to use hardware interrupts instead of polling to wait for events. The benefits of this are:
- it enables an O(1) implementation of "select" , instead of O(n), where n is the number of waiting objects
- it enables the CPU to go into a low power state when the event loop runs out of ready tasks
idle owl
@slender iron The fix is merged into the Blinka bleio library. I commented on the PR. If they need help with updating their branch from main, someone else will need to step in. That's a huge stretch for my Git fu.
tulip sleet
Agreed, there are two cases for async-ification: digitialio vs busio. Some busio...
My impression about Event is that the task awaiting the Event would not be scheduled until the event fired (however that happened, by interrupt, or polling). There is also ThreadSafeFlag (maybe it was ThreadSafeEvent earlier). The MicroPython discussion I linked to earlier discusses using ThreadSafeFlag in interrupt handlers, though I think they mean MicroPython handlers (both hard and soft interrrupts). But there could be a similar kind of thing implemented natively. That discussion describes waiting on Events or Flags as the paradigm. https://github.com/micropython/micropython/discussions/9131. MicroPython as a native select, (maybe as uselect) and I think it's even enabled for CircuitPy, but we don't use it that I know of.
lone sandalBOT
tulip sleet
@timid bolt we are planning to discuss native events and the like with the MPy folks, as I think I mentioned
timid bolt
<@1065334137243836517> we are planning to discuss native events and the like wit...
Sounds good. Let me know when that discussion will take place.
tulip sleet
it will be just us and them, but we can pass on questions
timid bolt
My impression about Event is that the task awaiting the Event would not be sched...
I've looked at the MP thread a few times, but I'm unsure what the takeaway is though. Ultimately there are many ways to do this and having a crisp definition of what the goals are would help guide decision. Whether it uses the Event class or not is not really the important part. The important part is how it will enable users to do more.
orchid basinBOT
still zephyr
@tulip sleet No luck with the TMP117 removing the CV class, I could do some testing removing some functionality. OR split in basic and advanced 🙂 https://github.com/adafruit/Adafruit_CircuitPython_TMP117/pull/13. let me know thanks 🙂
idle owl
<@329766224093249548> No luck with the TMP117 `removing` the CV class, I could d...
Dan's out for a bit today, not sure if he'll be around later this evening, or tomorrow. Simply letting you know so you have the right expectations on when you will receive a response.
still zephyr
Dan's out for a bit today, not sure if he'll be around later this evening, or to...
Thanks :). No worries
idle owl
You're welcome!
slender iron
@onyx hinge are you interested in the bangle.js 2 as well?
onyx hinge
@slender iron I have one.. but I don't really wear watches
slender iron
ah, ok. maybe you can help test then 🙂
blissful pollen
Use make V=1, make V=2 or set BUILD_VERBOSE similarly in your environment to increase build verbosity.
GEN build-itsybitsy_m4_express/genhdr/moduledefs.h
QSTR updated
File "<fstring>", line 1
(offstart=)
^
SyntaxError: invalid syntax
make: *** [../../py/py.mk:274: build-itsybitsy_m4_express/genhdr/compression.generated.h] Error 1
make: *** Waiting for unfinished jobs....```
Someone know offhand what tool I need to update (guessing?) to fix this? updated to the latest code after a while and failing. ATMEL and RP2040 ports triedIn case anyone else ever sees this error in the future the problem was python 3.7 was too old. Oops
wraith crow
Three people have posted about the fstring build error in the last month or two. Can the print(f"a={a}) statment in fstring.py be wrapped in a try/except block? I'm assuming if the program doesn't crash and the a=1 print message doesn't show up a more appropriate error is displayed.
never mind, it's a syntax error, try/except won't catch it :/
dense glacier
In case anyone else ever sees this error in the future the problem was python 3....
Kids and their fstrings these days lol. Thanks for the tip 👍
I didn't get to post a #circuitpython2023 in time but thought I'd follow up with a quick item to share...
I recently tried to mash up the asyncio and simple web server libraries. Unsuccessfully.
I would love something similar to the Python Tornado library in CircuitPython. MicroTornado if you will.
https://www.tornadoweb.org/en/stable/
Not sure if that is even the right solution today but thought I'd share.
Thank you all for your work! Hugs
tulip sleet
I didn't get to post a #circuitpython2023 in time but thought I'd follow up with...
I started working on async requests (not a server) here: https://github.com/dhalbert/Adafruit_CircuitPython_Requests/tree/async. That branch is now out of date, but somebody got it working (maybe with some changes -- that code was not even tested).
There are a number of MicroPython examples: https://www.google.com/search?q=asyncio+http+server+micropython&oq=asyncio+http+server+micropython
trim marsh
how can i install circuit python on NXP imx8 based board? , i have not found in the port folder
stuck elbow
it's possible it's not supported
tulip sleet
how can i install circuit python on NXP imx8 based board? , i have not found in...
imx8 are Cortex-A. We have not ported CircuitPython to these processors.
slender iron
@tulip sleet and @onyx hinge want to talk about rc.0 today?
tulip sleet
Sure - i am not making a lot of progress on the long-lived wifi bugs
i have a few things to check
slender iron
@tulip sleet we can always fix in another rc too
tulip sleet
i just meant I wanted to look at some stuff before the mtg, is all
don't have any known fixes right now
onyx hinge
@slender iron I just sat down to lunch, but can talk anywhere from 15 minutes to 4 hours from now
slender iron
15 min works for me
onyx hinge
@slender iron @tulip sleet I'll be ready to join a video call as soon as my cup of tea is ready
tulip sleet
i'm ready too
onyx hinge
I popped in to the "amelia" channel
tulip sleet
@slender iron we are in amelia
brazen hatch
Wow, circuitpython is y2k38 ready..
As I was setting the rtc I was expecting it to hard crash, but it just worked..
still zephyr
General Question regarding CircuitPython community Libraries, releases to Pypi were done automatically, however It seems that I have not update correctly my under the hood library files, as libraries are not longer updating automatically :(. any guidance would be appreciate: ex: https://pypi.org/project/circuitpython-styles/ vs https://github.com/jposada202020/CircuitPython_styles. Thanks in advance 😊Found the answer. needed to update release_pypi.yml and release_gh_yml in github workflows. 😊
tulip sleet
@onyx hinge @slender iron I finished the draft release notes for rc.0. I added four PR's to the 8.0.0 milestone: https://github.com/adafruit/circuitpython/milestone/30
**update frozen libraries **
#7491 opened 31 minutes ago by dhalbert
**Add frequency setting for RP2040 boards. **
#7430 opened 3 weeks ago by Lanzaa
Espressif ESP32-S3 DevKitC-1-N32R8 board New board or update to a single board
#5999 opened on Feb 9, 2022 by anecdata
**Add function for drawing polygons to bitmaptools **
#7471 opened last week by matemaciek
Of these #7491 is required. is #7430 done? #5999 was not working but would be nice to have. #7471 I think may be close to done? We can move some of these off the milestone depending on what you think.
onyx hinge
@tulip sleet hum about the frequency changing one .. if it breaks USB that's a pretty huge caveat! I hadn't thought about that, I just concentrated on the USB aspect of it.
onyx hinge
I guess the CI on 7491 was relatively quick because the build knows which boards have the frozen modules? have to admit that's pretty cool!
tulip sleet
I guess the CI on 7491 was relatively quick because the build knows which boards...
yeah, I agree -- I was surprised
proven garnet
@onyx hinge I saw you submitted this issue for the CI deprecation: https://github.com/jenschelkopf/issue-label-notification-action/issues/29
JavaScript isn't my first (or any) language, but I did manage to put together what we essentially need as an action
It needs a token with the proper permissions to write comments on issues, but otherwise works.
Happy to transfer over to Adafruit - I only own it now because I can't generate repositories for the org. Or just happy to own it, whatever is better/easier.
analog bridge
Would it make sense to change the esp32_camera module to espcamera so its more uniform with other esp specific modules?
stuck elbow
and rename the take method to be compatible with parallel capture
manic glacierBOT
thorny flax
import alarm library
manic glacierBOT
brazen hatch
I don't have time to actively work on code, however I could stress test espnow
I have 4 esps
Is it ready for testing?
analog bridge
I don't have time to actively work on code, however I could stress test espnow
I...
Yes, go ahead 
The example in espnow/__init__.c is wrong. Use the docs in espnow/ESPNow.c instead.
brazen hatch
alright thanks!
tulip sleet
Would it make sense to change the `esp32_camera` module to `espcamera` so its mo...
that makes sense, but it could be done as an alias now, and then in 9.0.0 we would drop the old name.
analog bridge
that makes sense, but it could be done as an alias now, and then in 9.0.0 we wou...
Is an alias necessary? esp32_camera is not in 7.x.x and is only in CP8 pre-releases.
tulip sleet
Is an alias necessary? `esp32_camera` is not in `7.x.x` and is only in `CP8` pre...
hmm, I didn't know that. @onyx hinge, what do you think, should we just change the name?
onyx hinge
I am fine changing the name if there is a more consistent choice. It does need to be espressif specific in some way though
tulip sleet
@analog bridge go ahead with a PR if you have time. I can get it into rc.0
onyx hinge
I chose the name because the underlying library is called https://github.com/espressif/esp32-camera but that doesn't have to match
tulip sleet
how about the take() name?
I don't think we have to be consistent with the library, since we may substitute some other library at some point
onyx hinge
I don't recall immediately why I chose that method name
(and I'm mobile right now so I can't investigate much)
tidal kiln
has anyone used transparency with ondiskbitmap? mainly curious how the desired color's index was found in the colorconverter
stuck elbow
I just use gimp, it lets you view the palette
tidal kiln
hmm. need to do it on the fly though. images will be grabbed from internet and stored locally. then brought in with odb.
stuck elbow
check what color the (0,0) pixel is?
or do a histogram and pick the color that has the most pixels
hmm, maybe only count the pixels on the edges?
there are many possible heuristics, but none is foolproof
tidal kiln
yah. no guarantee it'll be at a given location or dominate 😦
but do know it's "white"
stuck elbow
then just iterate over the palette and find the white
tidal kiln
that's what i'm doing with image_load
but obd doesn't provide a palette, instead it's a colorcoveter
stuck elbow
even for indexed images?
iirc colorconverter is used for 24 bit images
indexed images use palettes
so if you know it's white, just call make_transparent with 0xffffff?
tidal kiln
it's palette index, not hex value
tidal kiln
ah. ok. looks like obd adapts. was working yesterday with 24 bit images.
i've changed now to 8bit, since don't need that depth
>>> type(obd.pixel_shader)
<class 'Palette'>
>>>
with 8bit you will get a palette
tidal kiln
and now it's a palette
stuck elbow
yeah
manic glacierBOT
slender iron
@tulip sleet I made two small changes to the release notes draft
tulip sleet
np should I know?
tulip sleet
nope, that's weird, but the URL changes every time you update the draft, so if you accidentally edited an old draft, it might get confused
slender iron
looks like it saved ok
(moved the note about audio playback taking in filename and remove the RISC-V note on the ulp support)
tulip sleet
good, thanks
tulip sleet
i deferred some PR's; only one left is the espcamera one. @onyx hinge is #7492 OK by you/
I could publish rc.0 tonight or over the weekend. Like to do it in time for Anne to get it in to the newsletter.
tulip sleet
@onyx hinge I think sed -i might have gotten a workout
slender iron
I've got a 3 for 1 PR that I think could be 8.1. https://github.com/adafruit/circuitpython/compare/main...tannewt:banglejs2?expand=1
its got bangles js 2 support, jdi color memory display support and acep eink support
tulip sleet
i can branch 8.x.x after rc.0 and before we merge that
tulip sleet
thanks for the reminder. as usual I may encounter circuitpython-org update issues (that seems to happen often, especially just before major release 🙂 )
slender iron
np 🙂
brazen hatch
I have been using stackless for all this time, had nothing to report on rp2, s2. c3, nrf52
I suggest it get's included in 8.1 alpha 0 so it can be publically tested.
This needs exposure and time to be deemed stable.
Once these very stable platforms are good with it, we can begin enabling it on the rest of the platforms
slender iron
@brazen hatch make a PR and I'll mark it for 8.1
brazen hatch
I will switch it out of draft status
slender iron
k, thanks. I switched the milestone to 8.1
brazen hatch
I will also fix the one board that is failing tomorrow, it just had seperate settings and forgor to switch them
tulip sleet
@slender iron I am not going to branch until after I make the release and then do some smoke tests before publishing and publicizing the release.
manic glacierBOT
The only SAMD board I have is the Sparkfun SAMD51 Thing Plus which boots 8.0.0 so I'm wondering if the issue is caused by one of the PewPewM4 changes. Have you tried backing out these two changes?
wraith crow
@lone axle I'm glad it worked out for you but it really looks to me like the CIRCUIPTY_USB_CDC = 0 line wasn't added in the beta 2 build just moved up higher in the file....
Maybe there were additional changes to the USB_CDC logic on the board level
lone axle
<@382939733107408897> I'm glad it worked out for you but it really looks to me l...
That is true, in that specific diff it's just being moved to a different section of the file (into alpha order). But that line definitely seems to be what causes the serial connection to be disabled. I'm not sure if it's possible or how to tell which release came most imediately after #4333. That PR seems to be the one where it was first set to 0.
tulip sleet
I added console and data then??
lone axle
Although, maybe the location within the file can have some effect on it? I didn't test moving it back to the top, just removing it entirely (I made the assumption that value 1 would be default and enabled)
tulip sleet
The order makes no diff
lone axle
phew, I am glad for that 😅
wraith crow
I think there was probably another change somewhere, we were just very lucky someone decided to move it so it showed up on that diff 😁
analog bridge
@tulip sleet A suggestion for release notes:
- In-place firmware update (`dualbank`) capability may be disabled in favor of a larger CIRCUITPY drive.
+ In-place firmware update (`dualbank`) capability can now be disabled (default) at runtime in favor of a larger `CIRCUITPY` drive.
tulip sleet
<@329766224093249548> A suggestion for release notes:
```patch
- In-place firmwa...
It looks like CIRCUITPY_DUALBANK is off only on the 2MB boards? ports/espressif/mpconfigport.mk
analog bridge
It looks like CIRCUITPY_DUALBANK is off only on the 2MB boards? `ports/espressif...
Ya, 2MB isn't enough to have two ota partitions.
tulip sleet
Ya, `2MB` isn't enough to have two `ota` partitions.
I was wondering about the word “default” in your suggested, change in the release notes, since there is no universal default
I agree with the addition of the word “runtime “; that’s a good change
analog bridge
dualbank is disabled by default on new firmware load, in case of an existing filesystem, the filesystem's configuration is respected
tulip sleet
`dualbank` is disabled by default on new firmware load, in case of an existing f...
I was looking at https://github.com/adafruit/circuitpython/blob/f907f20eb98facb4d18e7cd9d2be79f1014d11bd/ports/espressif/mpconfigport.mk#L26, so on a new load, it is enabled, except on small flash boards, is that correct?
analog bridge
2MB:
disabled all the time, import dualbank will give ModuleNotFoundError
>2MB:
a runtime setting, import dualbank will give RuntimeError: storage is extended when disabled
dualbank is disabled by default on completely new firmware load, in case of an existing filesystem (e.g. uf2 update), the filesystem's configuration is respected
I feel like this needs to be documented, likely on the dualbank page
tulip sleet
I will look at your newwording, and maybe we need to update the doc too https://docs.circuitpython.org/en/latest/shared-bindings/dualbank/index.html because I think this is not really written down anywhere.
analog bridge
I will look at your newwording, and maybe we need to update the doc too <https:/...
I'll do a PR, also need to add a good example for dualbank
tulip sleet
I'll do a PR, also need to add a good example for `dualbank`
Thank you!
jaunty juniper
to be clear, the module is available on all boards (except 2MB boards) but the drive is now formatted in extended mode by default (thus the module is disabled until reformatting in non-extended)
right ?
disabled/enabled vs available/builtin/compiled in/whatever ?
analog bridge
@tulip sleet I updated #5999, can you do a re-test.
tulip sleet
<@329766224093249548> I updated #5999, can you do a re-test.
thanks for working on that
do you have that board?
analog bridge
to be clear, the module is available on all boards (except 2MB boards) but the d...
ya, that sums it up + any existing filesystem configuration will be respected, for e.g. on uf2 load, existing filesystem extension setting is kept
analog bridge
do you have that board?
no, but it should work now, opi flash needs to be flashed in dout mode
tulip sleet
so i need different args on the esptool.py command line
analog bridge
yes:
# Note: we use esptool.py to flash bootloader in
# dio mode for QIO/QOUT, bootloader then upgrades
# itself to quad mode during initialisation
config ESPTOOLPY_FLASHMODE
string
default "dio" if ESPTOOLPY_FLASHMODE_QIO
default "dio" if ESPTOOLPY_FLASHMODE_QOUT
default "dio" if ESPTOOLPY_FLASHMODE_DIO
default "dout" if ESPTOOLPY_FLASHMODE_DOUT
# The 1st and 2nd bootloader doesn't support opi mode,
# using fastrd instead. For now the ESPTOOL doesn't support
# fasted (see ESPTOOL-274), using dout instead. In ROM the flash mode
# information get from efuse, so don't care this dout choice.
default "dout" if ESPTOOLPY_FLASHMODE_OPI
so the web-based flasher may not work, I wonder?
analog bridge
I don't know about that but we should be defining these ESPTOOL settings in the sdkconfig as that is what multiple idf tools use
currently these settings are overridden by those defined in mpconfigboard.mk
tulip sleet
are the settings embedded in the .bin?
i am getting Warning: Image file at 0x0 is protected with a hash checksum, so not changing the flash mode setting. Use the --flash_mode=keep option instead of --flash_mode=dout in order to remove this warning, or use the --dont-append-digest option for the elf2image command in order to generate an image file without a hash checksum when trying to specify on the cmd line
if i don't specify, it loads but then boot loops
analog bridge
is this with a new firmware built with dout?
tulip sleet
ESP-ROM:esp32s3-20210327
Build:Mar 27 2021
rst:0xc (RTC_SW_CPU_RST),boot:0x8 (SPI_FAST_FLASH_BOOT)
Saved PC:0x4210b566
SPIWP:0xee
Octal Flash Mode Enabled
For OPI Flash, Use Default Flash Boot Mode
mode:SLOW_RD, clock div:2
load:0x3fce3808,len:0x68
load:0x403c9700,len:0x980
load:0x403cc700,len:0x2574
entry 0x403c98a4
I took the artifact from #5999
analog bridge
@tulip sleet can you try dio flash mode
tulip sleet
<@329766224093249548> can you try `dio` flash mode
I think I will need to build it myself, since I can't override the flash mode to esptool.py
(as implied by the error msg above??)
tulip sleet
<@329766224093249548> can you try `dio` flash mode
changed to dio in mpconfigboard.mk, still bootlooping, v similar to above (Saved PC is diff)
solid juniper
Can someone help me with creating a VID/PID pair for the Flipper Zero WiFi Dev Board?
solid juniper
Can someone help me with creating a VID/PID pair for the Flipper Zero WiFi Dev B...
I've forked the repo, but I'm not sure what folder to make in the org folder.
analog bridge
@tulip sleet flash this using:
CIRCUITPY_ESP_FLASH_MODE=dio
CIRCUITPY_ESP_FLASH_FREQ=80m
CIRCUITPY_ESP_FLASH_SIZE=32MB
bootloop:
SP-ROM:esp32s3-20210327
Build:Mar 27 2021
rst:0xc (RTC_SW_CPU_RST),boot:0x8 (SPI_FAST_FLASH_BOOT)
Saved PC:0x4210a89e
SPIWP:0xee
Octal Flash Mode Enabled
For OPI Flash, Use Default Flash Boot Mode
mode:SLOW_RD, clock div:1
load:0x3fce3808,len:0x68
load:0x403c9700,len:0x980
load:0x403cc700,len:0x25a0
entry 0x403c98a4
I am also getting messages:
WARNING: Flasher stub doesn't fully support flash size larger than 16MB, in case of failure use --no-stub. for both esptool.py 3.3.2 and 4.4
I think this PR has not been merged from main recently. There are esp-idf changes which may help this which are in the recent PR I did to update the esp-idf
analog bridge
I updated the PR to get the recent v4.4 commits
I am doing another build with octal flash enabled in sdkconfig.
@tulip sleet flash this using:
CIRCUITPY_ESP_FLASH_MODE=dout
CIRCUITPY_ESP_FLASH_FREQ=80m
CIRCUITPY_ESP_FLASH_SIZE=32MB
is it possible to build like this with some syntax ? Setting a define on build
make BOARD=raspberry_pi_pico CFLAGS="-DCIRCUITPY_SKIP_SAFE_MODE_WAIT"
that is only setting it in cpp; try make ... CIRCUITPY_SKIP_SAFE_MODE_WAIT=1. That will set it in make, and then the .mk files will define the CFLAGS version (assuming they are set up that way)
jaunty juniper
ah thanks, I don't remember if I tried that !
and I assume it's not possible to overwrite boardconfig.mk (since the CIRCUITPY_* are not set conditionally there)
tidal kiln
why doesn't OnDiskBitmap like this BMP? it works with adafruit_imageload, but ODB has a skipped columns appearance
Adafruit CircuitPython 7.3.3 on 2022-08-29; Adafruit Matrix Portal M4 with samd51j19
>>> from adafruit_matrixportal.matrixportal import MatrixPortal
>>> import displayio
>>> import adafruit_imageload
>>> mp = MatrixPortal()
>>> odb = displayio.OnDiskBitmap('image-formatter.bmp')
>>> tg1 = displayio.TileGrid(odb, pixel_shader=odb.pixel_shader)
>>> mp.splash.append(tg1)
>>> bmp, pal = adafruit_imageload.load('image-formatter.bmp')
>>> tg2 = displayio.TileGrid(bmp, pixel_shader=pal)
>>> mp.splash.append(tg2)
>>>
🎉
code.py output:
Hello World!
Code done running.
Adafruit CircuitPython 7.2.0-alpha.1-113-g401d160c5-dirty on 2023-01-28; ESP32-S3-DevKitC-1-N32R8 with ESP32S3
>>>
Finally... 🥳
tulip sleet
still wondering if we'll need special args on esptool.py, I'll try a plain flash with no args after the artifact builds
tulip sleet
ah thanks, I don't remember if I tried that !
and I assume it's not possible to ...
I think that's right
I just always edit the board file
or maybe it does force override... you could try it
but i don't think so
tulip sleet
I've forked the repo, but I'm not sure what folder to make in the `org` folder.
I don't know which repo you mean
analog bridge
Basically, the esptool settings in mpconfigboard.mk have to be removed for all boards, it just creates two sources of information and messes things up
tulip sleet
but we need them for the esptool.py flash etc. targets in Makefile??
the problem is not using cmake, maybe (??????)
i have to keep re-deriving how our build process works
i forget
solid juniper
I don't know which repo you mean
https://github.com/pidcodes/pidcodes.github.com is the repo. I'm following the instructions from https://pid.codes/howto/ to make the pair.
tulip sleet
https://github.com/pidcodes/pidcodes.github.com is the repo. I'm following the i...
what VID/PID is the Flipper using already?
analog bridge
but we need them for the esptool.py flash etc. targets in `Makefile`??
They can be inferred from sdkconfig with some parser script or maybe idf has some way to do it, I just want to have one true source of information.
Why does it have to be sdkconfig? Because it is known to idf as the only source of config information.
As for espressif_esp32s3_devkitc_1_n32r8. For now setting CONFIG_ESPTOOLPY_OCT_FLASH=y in board's sdkconfig should work with dout mode.
And yes the bin does contain these flash config settings
This can be viewed by using esptool.py image_info --version 2 *.bin. (available in idf v5.0)
tulip sleet
They can be inferred from `sdkconfig` with some parser script or maybe `idf` has...
agree with that, we can work on that. I'll retest when the rebuild is done. will be afk for a while in a little while. If it works I'll merge for rc.0
onyx hinge
wahoo excellent detective work you two
tulip sleet
i am just the tester 🙂
analog bridge
i am just hacking things together 🙂
solid juniper
what VID/PID is the Flipper using already?
It's using the Saola 1 w/WROVER pair at the moment, since that's the board I used as a base.
tulip sleet
@onyx hinge thank you in any case
lone axle
why doesn't OnDiskBitmap like this BMP? it works with adafruit_imageload, but OD...
That is weird. I don't know what is making it do that. I opened it in GIMP and re-exported without changing anything, the resulting bmp is a different size and when drawn by displayio doesn't have a different between odb and imageload
tulip sleet
It's using the `Saola 1 w/WROVER` pair at the moment, since that's the board I u...
i mean if you plugged in the Flipper now, what VID/PID's does it use?
without circuitpython
for its custom firmware or whatever
solid juniper
i mean if you plugged in the Flipper now, what VID/PID's does it use?
The Dev Board? I'm not sure how to check it.
tulip sleet
an actual Flipper, I eman
tidal kiln
That is weird. I don't know what is making it do that. I opened it in GIMP and r...
same here. both saving / not saving colorspace work when exporting from GIMP. maybe some clues here?
$ file image-formatter.bmp
image-formatter.bmp: PC bitmap, Windows 98/2000 and newer format, 64 x 32 x 8
$ file image-formatter2.bmp
image-formatter2.bmp: PC bitmap, Windows 3.x format, 64 x 32 x 8, image size 2048, resolution 2835 x 2835 px/m, 256 important colors, cbSize 3126, bits offset 1078
$ file image-formatter3.bmp
image-formatter3.bmp: PC bitmap, Windows 95/NT4 and newer format, 64 x 32 x 24
first is original that does not work
2 is exported from GIMP with "do not write color space" selected
3 is same but with option selected
both GIMP exported ones work
stuck elbow
The Dev Board? I'm not sure how to check it.
lsusb
lone axle
same here. both saving / not saving colorspace work when exporting from GIMP. ma...
interesting. The raw data diff shows a difference in the first part and then same data after that. (albeit, not so happily about reading binary data as text)
There must be some additional configuration or something in the header causing a difference.
jaunty juniper
ah CIRCUITPY_SKIP_SAFE_MODE_WAIT is a bad example, it's not in the makefile (there's literally a single occurence of it, the ifdef)
tulip sleet
ah `CIRCUITPY_SKIP_SAFE_MODE_WAIT` is a bad example, it's not in the makefile (t...
also we try not to use ifdef, instead using if so we can catch typos
tidal kiln
interesting. The raw data diff shows a difference in the first part and then sam...
yep. most likely something being mishandled in header.
wraith crow
When I plug my flipper in, Windows displays VID_0483&PID_5740 I don't have the WiFi Dev board though, that has it's own ESP chip I believe.
and USB port....
solid juniper
lsusb
How do I use lusb to see the Flipper's VID/PID pair? I'm porting CircuitPython to the WiFi Dev Board for, not the actual Flipper.
solid juniper
When I plug my flipper in, Windows displays VID_0483&PID_5740 I don't have the W...
I can check the Dev Board since I have one.
lone axle
Wikipedia has a decent looking listing of the header: https://en.wikipedia.org/wiki/BMP_file_format#Bitmap_file_header I've never delved into the details of it that far before, but I'm curious about what could be causing it, next week I'll try to compare imageload code to the core code inside of ODB to see if I can spot anything that would account for it.
stuck elbow
How do I use `lusb` to see the Flipper's VID/PID pair? I'm porting CircuitPython...
you connect it, and you type "lsusb" in the linux terminal, and press enter, it will list vid and pid numbers of all devices connected to usb
tulip sleet
0483 / 5720 is STM's "Mass Storage Device" VID/PID
wraith crow
The flipper does use an STM chip in the main unit
solid juniper
you connect it, and you type "lsusb" in the linux terminal, and press enter, it ...
Would you like me to use the firmware that the Dev Board is shipped with?
stuck elbow
as long as it has a usable vid/pid that you could reuse
solid juniper
as long as it has a usable vid/pid that you could reuse
Alright. I'll re-flash the blackmagic firmware and check.
tidal kiln
Wikipedia has a decent looking listing of the header: https://en.wikipedia.org/w...
cool. thanks! please ping me with any updates.
lone axle
cool. thanks! please ping me with any updates.
Will do. Here is a slightly more helpful way to visualize the difference.
stuck elbow
this is why I switched to png as soon as I figured out how to read it
analog bridge
@tulip sleet I pushed a commit enabling octal flash in sdkconfig
https://github.com/adafruit/circuitpython/pull/5999/commits/858e212e34f9ffaaa1c2b5b17331e52bf780b6ad
jaunty juniper
also we try not to use ifdef, instead using `if` so we can catch typos
ok I added it to circuitpy_mpconfig.mk (and used #if), and it works now. Good to know. (And I can PR that too)
tulip sleet
ok I added it to `circuitpy_mpconfig.mk` (and used #if), and it works now. Good ...
👍 they are in alphabetical order, if you did not already do that
solid juniper
as long as it has a usable vid/pid that you could reuse
303a/4001 is the VID/PID pair for the Blackmagic firmware.
tulip sleet
303a/4001 is the VID/PID pair for the Blackmagic firmware.
303a is Espressif's VID. Espressif allocates PIDs to customers here: https://github.com/espressif/usb-pids. But 4001 is Espressif's own PID -- the BlackMagic sw is just reusing that. Ideally the Flipper's developers would submit a PR to that repo to get a PID. You are not an official rep of Flipper, so you can't do that.
pid.codes would not be involved here
solid juniper
303a is Espressif's VID. Espressif allocates PIDs to customers here: <https://gi...
I guess I'll head to them for that. Thanks.
tulip sleet
I guess I'll head to them for that. Thanks.
Espressif might give it out to you, example: https://github.com/adafruit/circuitpython/pull/6079.
if the wifi module were essentially just a module on a board with no changes, then you could just use the module PID. But you want to add a new board here, right, so there's something different you want to include that does not match an existing board
so what is extra/different on this board?
tulip sleet
@solid juniper see last post above ^^
timid bolt
Is there a good way to allocate aligned memory in CP/MP?
tulip sleet
Is there a good way to allocate aligned memory in CP/MP?
Heap blocks are 16 bytes each, aligned
timid bolt
Yeah, any bigger than that though? For a ring buffer aligned to its size.
tulip sleet
Yeah, any bigger than that though? For a ring buffer aligned to its size.
I don't think there's any API for that. What needs that alignment?
timid bolt
You can setup rp2 dma to work on a ring buffer. It's more efficient if you can use a ring buffer to stream data to/from a peripheral.
tulip sleet
but it has to be aligned to the buffer size, hmm?
timid bolt
yes, for the pointer arithmetic to work
tulip sleet
i have no answer, sorry, I don't think we use that technique anywhere .
solid juniper
so what is extra/different on this board?
I've set an inverted PWM RGB LED in CircuitPython, since the board's RGB LED is 3 LEDS (Blue to IO4, Green to IO5, and Red to IO6). IO3 isn't broken out, but it is an I2C Pull Up pin. IO45 and IO46 aren't used or broken out, and go to GND. IO0 isn't broken out, but the BOOT button is attached to it.
timid bolt
No worries. There's an old trick where you can allocated twice as much memory and find your aligned chunk inside it. But seems too wasteful for a limited resource microcontroller.
tulip sleet
I've set an inverted PWM RGB LED in CircuitPython, since the board's RGB LED is ...
got it, thanks, so would make sense to ask Espressif directly or indirectly for a CircuitPython PID
solid juniper
got it, thanks, so would make sense to ask Espressif directly or indirectly for ...
I can make a PR on their repo.
tulip sleet
Lots of examples of getting PIDs for other boards: https://github.com/adafruit/circuitpython/pulls?q=is%3Apr+"espressif%2Fusb-pids"
jaunty juniper
there's a #ifdef CIRCUITPY_ULAB I noticed that would always be true, wouldn't it ?
https://github.com/adafruit/circuitpython/blob/cc822e0d8d4fd7af22b403835e375c79421ac693/shared-bindings/adafruit_pixelbuf/PixelBuf.c#L43
tulip sleet
there's a `#ifdef CIRCUITPY_ULAB` I noticed that would always be true, wouldn't ...
Yup, that's wrong, though I assume it has no effect, since it's just an #include.
atomic summit
woops! Anyone know the status os Venture 13.2 and UF2 ? I'm staying on 13.1... but I have a user reporting he cant copy over CPY firmware onto his ProS3 UFT drive... like what was happening in 13.01
tulip sleet
woops! Anyone know the status os Venture 13.2 and UF2 ? I'm staying on 13.1.....
we haven't heard this yet, that's too bad if true. I can upgrade one machine to 13.2 and see what happens
atomic summit
we haven't heard this yet, that's too bad if true. I can upgrade one machine to ...
Ok, thanks.. I only have the 1 Mac atm, and cant risk moving to 13.2 if it's busted again 😦
tulip sleet
is it a UF2 or a .hex that is not uploading properly/
atomic summit
.uf2 CPY file
tulip sleet
It is also still true that copying .hex files larger than 1MB does not work (e.g., on the micro:bit). See https://github.com/ARMmbed/DAPLink/issues/982 for workarounds and further investigation.
atomic summit
He tried dragging and copy -X
tulip sleet
I'll try, and complain right away if it's broken again 😦
atomic summit
I'm working with the guy to re-flash everything from scratch
tulip sleet
I'm working with the guy to re-flash everything from scratch
a Mac is all he has? no RPi lying around?
atomic summit
he has a Linux install, and he's also having issues there... so though he is getting into UF2 bootloader mode, I think he's trashed his flash putting things in the wrong place.
tulip sleet
I am updating but this is a 2017 Macbook, will be a while
atomic summit
I am updating but this is a 2017 Macbook, will be a while
Ok, I'll report back once I have had him erase and re-flash
Yeah, false alarm... an erase_flash and re-flash of the UF2 bootloader got the board back into a usable state and he was able to copy CPY over wit drag and drop! few!
Sorry @tulip sleet
tulip sleet
whew, well, I'll upgrade and test anyway, may as well. No problem.
tulip sleet
Yeah, false alarm... an erase_flash and re-flash of the UF2 bootloader got the...
Confirming: dragging UF2 works fine on Ventura 13.2.
atomic summit
Confirming: dragging UF2 works fine on Ventura 13.2.
Woohoo! Yay Apple 😉
Thanks very much for confirming !
brazen hatch
Why do you have esp32c3 so much?
Me:
Preparing the issue now.
It's so easy to reproduce
TLDR; serial is useless
jaunty juniper
Why do you have esp32c3 so much?
Me:
does that say beta 5 ?
brazen hatch
yep, I did flash latest stackless onto it now though
It has been like that since forever, it's not version specific.
Tested stackless and now threw back in the box of cursed electronics.
wraith crow
Tested stackless and now threw back in the box of cursed electronics.
I ended up writing a serial transfer program to copy files character by character checking that each one made it correctly to get things on the C3. If I tried to transmit more than one character at a time I ran into problems. Once the code is on the C3 the serial issue isn't as frustrating, occasionally commands typed in the REPL drop a character but probably only when I'm typing fast 😁. I keep one of the C3s on my network with a static IP address just so I can bookmark the webpage as a jumping off point to find dynamic IP addresses of whatever boards I'm working on. It also runs a neopixel color wheel just for fun 😁 At some point if you don't beat me to it, I'll try and nail down a simple script demonstrating the problem and open an issue.
When I first got the C3 it was totally unusable, the REPL was insane so I was very happy when it got some attention during the web workflow rollout and at least became usable.
brazen hatch
the moment you throw something like jcurses onto it and do ansi escape stuff, it becomes a meme
brazen hatch
Doing proper tags once in a while is refreshing.
You can now sit and relax enjoying the cool repl.
I remember when I was doing the picow-ap I had made it to -150..
orchid basinBOT
manic glacierBOT
tulip sleet
manic glacierBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
In the BLE Workflow, CIRCUITPY_BLE_NAME in settings.toml allows definition of the advertised name. This is useful when multiple boards are within range.
When multiple boards with web workflow enabled on the same network, it would be useful to have an equivalent CIRCUITPY_WEB_NAME.
lone axle
Does anyone have insight into the IS31FL3741 library / core module development? I'm curious whether the core module class documented here: https://docs.circuitpython.org/en/latest/shared-bindings/is31fl3741/index.html#is31fl3741.IS31FL3741 (not available on very many devices).
Is intended to have the same functionality as the python implementation of a class with the same name inside the library. It's documented here: https://docs.circuitpython.org/projects/is31fl3741/en/latest/api.html#adafruit_is31fl3741.IS31FL3741
They seem to have similar functionality, but perhaps not 100% the same. I don't know whether that is intentional or not though, or whether it would make sense to try to modify one or the other to make them match so one is a drop in replacement for the other.
turbid radish
Who was the author? Scot, or maybe Ladyada?
manic glacierBOT
lone axle
On the library side the absolute oldest commits are from Ladyada, but several of the early ones are from Paint Your Dragon. I'm not sure about the core side of things.
slender iron
@lone axle I think gamblor did the core side
lone axle
Uh oh, todays status file for the meeting seems to have had trouble generating: https://adafruit-circuit-python.s3.amazonaws.com/adabot/bin/reports/circuitpython_library_report_20230130.txt
Yesterday's is good, so I am going to copy from there for the notes doc today.
manic glacierBOT
proven garnet
Uh oh, todays status file for the meeting seems to have had trouble generating: ...
Oh! I came to bring this up!
I triggered a rerun, not sure if it was a fluke failure or something systemic but I don't think anything was change in adabot to trigger this.
It may generate a new report in time, but worst case it'll help diagnose whether it's an actual issue.
lone axle
Thank you!
proven garnet
@lone axle Looks like it should generate in time and that it was a fluke failure - want me to update the document when it's done?
lone axle
<@382939733107408897> Looks like it should generate in time and that it was a fl...
Sure, that would be great. Thank you.
tulip sleet
@lone axle are you running the mtg or kattni (it's just that you're top of the list in the round-robins, so I wasn't sure)
proven garnet
Updated the doc
lone axle
<@382939733107408897> are you running the mtg or kattni (it's just that you're t...
I am running today. Kattni asked me to cover for today.
timid bolt
Is there a document for the meeting?
ornate breach
idle owl
@timid bolt ^^ (should have replied to your message)
timid bolt
Is there are standard place to find the docs, or do we just wait for the link here?
proven garnet
A message with the link for the relevant (active meeting or upcoming meeting) document is pinned to this channel
timid bolt
A message with the link for the relevant (active meeting or upcoming meeting) do...
thank you
lone axle
<@&356864093652516868> the weekly meeting will begin here on discord in just under 30 minutes from now. The notes doc is here https://docs.google.com/document/d/1kv7EzMdwVLd8ODEcqrHhSPrz_MGrS5Q9qMI6CgXBbXg/edit?usp=sharing if you'd like to add your status updates and hug reports to it. We look forward to seeing everyone during the meeting!
proven garnet
thank you
You can find old meeting notes at https://github.com/adafruit-circuitpython-weekly-meeting
idle owl
Hey folks - we're finishing up our internal meeting. Might be a couple of minutes over.
ornate breach
Working on a smöl TFT bff for the Qt Py https://mastodon.social/@Oakdevtech/109764516959695403
More small circuitpython displays 🙂
still zephyr
🙂
storm minnow
Working on a smöl TFT bff for the Qt Py https://mastodon.social/@Oakdevtech/1097...
What if all the pins were near one of the ends?
ornate breach
What if all the pins were near one of the ends?
They could, but I figured it would be better if it was centered.
onyx hinge
@ornate breach is it designed to go "under", like the neopixel 5x5 bff is?
ornate breach
Yeah
onyx hinge
I also wondered a little bit about interference with the usb connector but it's probably fine
storm minnow
They could, but I figured it would be better if it was centered.
Hmm… I just feel like having the USB uncovered would be good, so you can more easily grip the cable. Definitely not gonna get in the way of the connector itself tho
ornate breach
If done right, you could use a battery bff and have a tiny mobile display
storm minnow
Maybe two versions? :0
ornate breach
🤷♂️
The display components mean you need to add a spacer
So USB shouldn’t be an issue
storm minnow
Just might be a little harder to grip the cable end to pull out
Since you’d need to grab the sides instead of the top and bottom
ornate breach
There shouldn’t be any issue but I’ll try out when ever I get the PCB and parts.
storm minnow
I don’t think it’ll be a problem problem, just maybe less than ideal
onyx hinge
@timid bolt in the Status Updates section, I moved your report so it's in alphabetical order, in case you are looking for it
blissful pollen
@lone axle Feel free to ping me about the IS31FL3741. I spent a lot of time in that and have done a bit of work getting it running on non-standard displays/boards too.
lone axle
Will do, thank you!
timid bolt
<@1065334137243836517> in the Status Updates section, I moved your report so it'...
Ahh, sorry, I didn't know that was the convention.
idle owl
<@382939733107408897> Feel free to ping me about the IS31FL3741. I spent a lot o...
@lone axle I feel like my wife did some work on making them work with the LED Animation library. I don't know if it's still sitting around or not, but I can check if you're interested.
idle owl
Ahh, sorry, I didn't know that was the convention.
No worries! We didn't mention it!
onyx hinge
@timid bolt the idea it's the host first, then everyone else in alphabetical order. but whatever's in the document, we roll with it
lone axle
<@382939733107408897> I feel like my wife did some work on making them work with...
Yep, there is a helper class that Rose made that adapts the glasses to work with led_animation. I think it can become the basis for one for the matrix as well.
blissful pollen
<@382939733107408897> I feel like my wife did some work on making them work with...
Pretty certain I had the animations working based on her work too. Via python or core implementations. Have to go look at my original code again
idle owl
Yep, there is a helper class that Rose made that adapts the glasses to work with...
Ah ok, I didn't realise we pushed it. Huzzah!
idle owl
Pretty certain I had the animations working based on her work too. Via python or...
Nice ok.
slender iron
👍
onyx hinge
[my colleague just informed me that I failed at alphabetical order 🤣 back to school for me]
thorny jay
Wait @onyx hinge , I would like a CP board to be a SNES mouse as I could not find one and I have a physical game that requires that...
Wondering how hard it would be the opposite of what you are doing, pretending to be a legacy hardware? Like SNES joystick (or mouse), or PS2 keyboard or ...
idle owl
For sure!
storm minnow
Emulating a peripheral sounds easier than emulating a whole console
turbid radish
YAY, thanks!!
onyx hinge
my gut tells me that rp2040 with pio should be a good platform for making an nes/snes controller emulator but as I don't have either I'm not going to dip my toes in
You "just" need to repeatedly load and scan out an 8- or 16-bit value from the microcontroller, according to the incoming clock and strobe signals.
and of course promptly update the values which will be one challenge
slender iron
definitely sounds doable with pio
onyx hinge
for joypad style just repeating the old value when none is available should be fine; for mouse you'd need to repeat a packet with no additional X or Y movement reported. It should all be possible though
still zephyr
Could the example in the community libraries be the benchmark so others folks could update their work. https://github.com/tannewt/CircuitPython_Example/tree/4a55a4a6d820ac80d511ee7ccc28f9e5b207e189. That way we could track the updates done.
proven garnet
I agree that documentation is far more feasible and digestible as well from the user perspective.
Awesome, thanks y'all!
That was essentially all of my wrap up, so big fan (at least a specific corner of tooling).
ember iris
I am curious at exploring live plotting and looking to create a matplotlib like library, but that's mostly the kind of thing I want to see this year--anything that makes it easier to 'see if the circuit is working' kind of thing (I'm planning on trying to make a plotting library this year, so who knows if it'll actually happen)
turbid radish
ChatGPT could do it
ember iris
CircuitGPT
crimson ferry
perhaps a minimized PR-type CI that users could use to build on Github, but just for one board
still zephyr
I am curious at exploring live plotting and looking to create a matplotlib like ...
ping me if you need anything with that. I love the idea 🙂
onyx hinge
yeah and expensive if you can't do it on github/microsoft's dime
ember iris
ping me if you need anything with that. I love the idea 🙂
I haven't finished writing my Circuit Python 2023 (which is fine to me, I saw it as more an exercise for me to roadmap what I want to work on this year) so this is my "I need to say it outloud to commit to this project" thing, I will definitely ask for some help once I spin it up
still zephyr
I haven't finished writing my Circuit Python 2023 (which is fine to me, I saw it...
🙂 no worries.
idle owl
Next meeting is Monday February 6, 2023, at 2pmET/11amPT!
still zephyr
Cheers
slender iron
thanks for hosting!
proven garnet
Thanks!
gilded cradle
Thanks
ember iris
Thank you all, have a great week!
onyx hinge
@blissful pollen circuitpython file "LICENSE_MicroPython" says that asf4 and mynewt-nimble are already Apache-2.0 licensed items. I did not verify this specifically but if accurate it would mean that we've already got Apache 2.0 licensed stuff in the mix
blissful pollen
<@228322543507865601> circuitpython file "LICENSE_MicroPython" says that asf4 an...
Thanks, I'll cross reference those as well. My other option was to e-mail the author if need be and ask them. Hopefully not needed.
lone axle
Here is the notes document for next Monday’s CircuitPython Weekly meeting. It is at the normal time of 11am Pacific / 2pm Eastern here on Discord. Everyone is encouraged to attend! Please add your hug reports and status updates even if you’ll be attending the meeting - it’s super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and we’ll read them off during the meeting. Hope to see you there! <@&356864093652516868>
https://docs.google.com/document/d/1BAM7-G2YJIgL8ptZG8rY5rKig1ptj1nfbDeuzX6dQzA/edit?usp=sharing
lone axle
@blissful pollen do you know if this python library class: https://docs.circuitpython.org/projects/is31fl3741/en/latest/api.html#adafruit_is31fl3741.IS31FL3741 is meant to be the same as this one implemented in the core module? https://docs.circuitpython.org/en/latest/shared-bindings/is31fl3741/index.html#is31fl3741.IS31FL3741 Or really the question is do they provide different functionality?
To me it looks like two different implementations with mostly the same functionality, but minor differences in the API.
There is code like this PixelBuf wrapper that seems to want to make use of the core module and currently doesn't work with the python library one. https://github.com/adafruit/Adafruit_CircuitPython_IS31FL3741/blob/main/adafruit_is31fl3741/is31fl3741_pixelbuf.py
If those two classes are ultimately providing the same functionalities we might benefit by equalizing the APIs so that either can be used in place of each other (with the core one being likely faster, but not available on all devices.)
proven garnet
@gilded cradle PlatformDetect had a botched release caused by a fault in my previous PR, fix submitted.
Oh you already have an issue for it
brazen hatch
storage.erase_filesystem() hanged on picow once. (I waited 3 minutes)
Powercycling the board, it was not erased.
Doing it again worked.
Weird.
blissful pollen
<@228322543507865601> do you know if this python library class: https://docs.cir...
The core was meant to be close the python (as the python class was written first). But they both are basically the driver to the chip at the lowest level that the PixelBuf/DisplayIO type classes can use to access the chip.
Same thing with the _Pixelbuf.py and led_glasses_animation.py. The pixelbuf is meant to work with the core changes I made vs the python only. The speed is the main difference.
Also as of now the IS31FL3741 core code is only compiled in the glasses board, though I have tried it successfully on other boards compiling it in manually.
jaunty juniper
the IS31FL3741_pixelbuf calls self.is31fl3741.write(self.mapping, buffer) from _transmit, but the python class doesn't have a write(), we had to add it to make it work
lone axle
The core was meant to be close the python (as the python class was written first...
Ahh, thank you! I'm going to try tinkering with the them a bit to see if it's possible that we can get it to where led_animations can work the same whether you use the core or the python class.
Yeah, I'm just about to jump back in, and was thinking I would try adding write into that class itself rather than in user code. To try to have the API match closer the core one.
jaunty juniper
yep
blissful pollen
Hmmm I never did upload my full code for the "sports team glasses" to github, but I can find it later if you want an example. I think I used the animation library but maybe I'm forgetting what I did
lone axle
Definitely happy to look at all the examples I can. Thank you.
tidal kiln
@jaunty juniper @tulip sleet thanks for this PR/review. just seeing this. also seeing some guide feedback chatter about this. but looks resolved.
https://github.com/adafruit/Adafruit_Learning_System_Guides/pull/2372
blissful pollen
Definitely happy to look at all the examples I can. Thank you.
tulip sleet
@slender iron @onyx hinge I did not bother to branch 8.0.x yet because there might be a couple more things that would be convenient to get into both 8.0.0 and main with single PR's before rc.2. But if you want to plow ahead on 8.1.0, I can branch.
lone axle
Thank you!
slender iron
I figured we would when something for 8.1 is ready to merge
manic glacierBOT
tulip sleet
sounds good, #7509 and #7510 could go in 8.0.0
onyx hinge
<@252717193496756235> <@426200547024961539> I did not bother to branch 8.0.x yet...
I'm not in a hurry at this moment
naive palm
Hi, is there any way to use the python-statemachine module in circuitpython?
tulip sleet
Hi, is there any way to use the python-statemachine module in circuitpython?
do you mean https://pypi.org/project/python-statemachine/ ? It looks like pretty normal python code. Only issue might be its size on various boards. You could just copy it over and try it
slender iron
anyone know why actions aren't running on this new library repo? https://github.com/adafruit/Adafruit_CircuitPython_SPD1656/pull/1
heh, and now it goes
tulip sleet
"Traffic is heavy this afternoon on GitHub CI; we see a clog near ubuntu-latest from the helicopter"
naive palm
do you mean <https://pypi.org/project/python-statemachine/> ? It looks like pret...
will do, thx
tulip sleet
also if it had a lot of dependencies, but I didn't see that with a quick look
orchid basinBOT
lone axle
@blissful pollen if you have a chance can you you try running the "native" example from the library repo examples folder? I'm finding that the native example seems to have broken behavior on 8.0.0, but is working correctly on 7.3.3. Hoping to get confirmation of that observation from someone else who has the hardware to rule out a fluke or other hardware issue on my end. I've made an issue to track it here: https://github.com/adafruit/circuitpython/issues/7516
orchid basinBOT
blissful pollen
<@228322543507865601> if you have a chance can you you try running the "native" ...
I'll give it a shot. Likely cannot get at it until tomorrow evening. Feel free to poke at me if you don't hear after that
timid bolt
If I have a digitalio pin in "open drain" mode then it needs to be pulled high somehow. The rp2 code does not use the on-chip pull-up, should it? Or does "open drain" mode expect to use external pull-ups?
stuck elbow
If I have a digitalio pin in "open drain" mode then it needs to be pulled high s...
you need an external pullup, the internal ones are way too weak
warm stump
Got my Seesaw, the little ATTiny817 looks to be working really well. It threw me for a second as it uses UPDI instead of SPI/ICSP or SWD, but Atmel Studio recognises UPDI so it was all good.
Made a little getting started source package which might need a bit of verification from Microchip
and a few additional resources like a Fritzing diagram but it is just 3 wires instead of 6 + StemmaQT
along with a bit of circuitpython busio code to use it as a raw I2C device once programmed over UPDI. This is additional to and alongside anything from the seesaw library, linked with web shortcuts.
this and a bit of 433MHz Doorbell Magic with the receiver from FS1000A, supposedly made by SSS Industrial Doors but quite common radio modules.
it may have some intermodulation with the US version of LoRa which uses a similar frequency, and also the UK version which is double. I called it ShoRa because the range was only 1-2m on low battery.
onyx hinge
do we use isort in our pre-commit? maybe not. If we do, we may be needing an update due to some incompatible change that occurred within the last week: https://stackoverflow.com/questions/75269700/pre-commit-fails-to-install-isort-5-11-4-with-error-runtimeerror-the-poetry-co
proven garnet
do we use isort in our pre-commit? maybe not. If we do, we may be needing an up...
We don't, though I've used it personally and was debating exploring introducing it.
I can't think of any repo that uses it that I've encountered, but good to know!
Also, how would I find the block size, block count, etc. of the file system of the pico running CircuitPython. Looking at http://github.com/wokwi/rp2040js but would need that info. Is it the same as MicroPython?
onyx hinge
kinda 😦 😦 about pre-commit this morning, can't figure out why it's throwing these messages in my personal project, but only when running under python 3.11!:
src/wwvb/updateiers.py:36:3: E1101: Module 'os' has no 'path' member (no-member)
src/wwvb/updateiers.py:58:28: E1101: Module 'io' has no 'StringIO' member (no-member)
they moved those around to keep people on their toes
tulip sleet
@timid bolt are you also "gneverov" on github, or is that someone else? I was going to ping you in a reply to https://github.com/adafruit/circuitpython/issues/7522
onyx hinge
I closed the 7.3.x milestone on github
timid bolt
you need an external pullup, the internal ones are way too weak
Being "too weak" depends on the application though. If the internal ones are very weak, then it probably doesn't hurt to turn them on anyway to avoid weirdness like:
>>> pin.value
False```It is me. I am fortunate to have a near globally unique name. 😁
timid bolt
FYI, I'm working on a proof of concept that allows native libraries to schedule a callback on the main loop from an ISR, and it does this without relying on an asyncio module implementation being in the core.
My hope is it will be acceptable for all relevant the core libraries to expose wait methods, with typically a small amount of C code, but not necessarily including an async module to save code size if you don't want it. The "wait" methods would throw an exception if there is no main loop set.
Returning an "awaitable" from C code is not in and of itself a problem. You can just go return m_obj_new(foo_t) where foo_t is a type that implements the awaitable protocol.
stuck elbow
Being "too weak" depends on the application though. If the internal ones are ver...
it does hurt, because then you don't get the warning about missing pull-ups, and you wonder why nothing is working
timid bolt
Where is this warning?
stuck elbow
Where is this warning?
when you create an i2c bus
timid bolt
Is busio implemented using digitialio? I thought it would be a native implemenation.
stuck elbow
bitbangio is implemented using the same functions as digitalio
but the point is, you never want internal pull-ups with open-drain pins
and if you would enable them by default, they would throw off the calculations of external pull-ups that you need anyways
jaunty juniper
busio is hardware busses, bitbangio is software bus.
busio.I2C tests the pull ups by counting how long it takes for the signal to come back up
stuck elbow
they would also sometimes randomly work, leading you to a wild goose chase about why things are not working depending on the phase of the moon
you generally want things to fail as early and as consistently as possible
timid bolt
I see. Makes sense in hindsight. Thanks. I might be helpful to mention the need for external pull-ups in the open drain drive mode documentation.
idle owl
It is me. I am fortunate to have a near globally unique name. 😁
I know that feeling! 😄
ornate breach
I know that feeling! 😄
Is Kattni short for anything or just an internet nickname you picked up in the olden days of the inter webs ?
For the longest time I went by UBRz and I still use NotUBRz to an extent these days
idle owl
Is Kattni short for anything or just an internet nickname you picked up in the o...
It was given to me by a friend I worked with decades ago by him writing me a note and addressing it to Kattni, and when I went back to uni after two years off, I decided it was time for a new name, and after trying out nicknames for my legal name, but none of them spoke to me. Then I remembered Kattni, and that was that.
ornate breach
It was given to me by a friend I worked with decades ago by him writing me a not...
I love that, thank you for sharing that anecdote. Finding a good name when the one you were given doesn’t fit is hard.
idle owl
I love that, thank you for sharing that anecdote. Finding a good name when the o...
You're entirely welcome. And fully agree.
@tulip sleet When you're around, I have a question/request. Ping me please!
ornate breach
For the longest time I didn’t like my first name of Seth. I got bullied a lot so I didn’t really want to go by that on the internet.
storm minnow
It was given to me by a friend I worked with decades ago by him writing me a not...
My name was given to me too :3 I’m curious tho, was Kattni a misspelling or just a nickname he made up?
ornate breach
Eventually I was using skerr for university related accounts so I decided it was close enough but not obvious.
idle owl
My name was given to me too :3 I’m curious tho, was Kattni a misspelling or just...
He made it up. 🙂
storm minnow
He made it up. 🙂
Hehe. Benefit of a made up name is you’re not likely to run into others using it!
idle owl
Hehe. Benefit of a made up name is you’re not likely to run into others using it...
Indeed. As far as I know, I'm basically the only one. Katni is a city in India, if I remember correctly. There's a "kattni" on Deviant Art that isn't me. Otherwise, except for names that include "kattni", the rest of them are probably me.
storm minnow
Indeed. As far as I know, I'm basically the only one. Katni is a city in India, ...
Nice. There are people with various spellings of Kyoshii, but mine appears to be unique
tulip sleet
<@329766224093249548> When you're around, I have a question/request. Ping me ple...
back!
idle owl
@tulip sleet Welcome back. I was informed that the Reverse TFT Feather board def was updated to include the buttons, but the update is not in a release. I can't link to "latest" for the guide. Is it possible to do another RC tomorrowish? Probably by end of week, I guess.
tulip sleet
Yes, I plan to do another one soon. Only one outstanding right now is https://github.com/adafruit/circuitpython/pull/7521
I can probably actually fix that up myself and merge it. When are you publishing the guide? ASAP?
idle owl
I can probably actually fix that up myself and merge it. When are you publishing...
It's a full new micro guide, so it'll take a bit I suppose. ASAP, but that's probably not until the end of the week at this point.