I would say don't consider our work as sacred or perfect in some way; there have been bugs in the past and there may be in the future.
#circuitpython-dev
tulip sleet
idle owl
Fair enough! Whatever makes things go faster, within reason, I'm up for trying.
proven garnet
How would I be able to figure out if a specific port's pin has OPEN_DRAIN available on a given pin as a drive mode? Is that defined somewhere I could see?
If it's complicated, I'm looking at the STM32F405 here but curious about a general case as well
Was able to find it in the data sheet, but still curious anyways!
crimson ferry
are there pins that don't have open drain?
proven garnet
I honestly have no idea, I know very very little about the inner workings electrically
crimson ferry
it can be fudged in code I believe if the chip doesn't natively support it
proven garnet
it can be fudged in code I believe if the chip doesn't natively support it
I see the I2C bitbang does this I believe
crimson ferry
in circuitpython, could write a pin finder to loop through all pins and try
tulip sleet
@onyx hinge I did not have time to look at the GPIO0 Feather RP2040 issue yet, but I wrote it up very quickly as the issue above. Going for a walk now. ^^
idle owl
Hmm. Is it still worth including a warning about "You must use CP 7.1.0 or later" in this asyncio template page? Or are we far enough past 7.1.0 that it doesn't need to be included anymore?
tulip sleet
i think we could remove it. 7.1.0 is last fall.
idle owl
Also, with asyncio being a library, how can I verify whether a board supports asyncio? (SAMD21 is obvious, none of them do.) This question is basically specific to the few nRF boards that evidently do not support it. Do I need to try it on those boards to find out if it works? Or is there some where to look it up?
idle owl
i think we could remove it. 7.1.0 is last fall.
Thanks, will do.
tulip sleet
Also, with asyncio being a library, how can I verify whether a board supports as...
I am not sure but I think if _asyncio is supported, then async/await are also turned on. Hmm, but _asyncio is not showing up in the module support matrix
idle owl
I am not sure but I think if _asyncio is supported, then async/await are also tu...
Yeah that's what I checked first.
tulip sleet
_asyncio and async/await don't have to go together, but I think they do now.
solar whale
Since there are a lot of boards that won't work -- would it be a good practice to use try/except on the import.
idle owl
Since there are a lot of boards that won't work -- would it be a good practice t...
Well, this template is designed to go into each board guide. So you simply wouldn't add it if it doesn't work with a particular board.
tulip sleet
this is the case of a language feature being on or off. The same is true of, say, long integers. Perhaps we should have a table for that.
Language Feature support matrix
idle owl
I mean for now I'll probably be the only one using this template. So I can deal with it not containing instructions to find out whether a board is supported or not.
But if it goes beyond me, there needs to be a way to determine whether it should go into a guide or not.
Other than SAMD21
idle owl
@tulip sleet Is it ok to use the word "simultaneous" when referring to the animations running on the rings? The actual sentence is this: You will use a button to control two simultaneous animations displayed on two different NeoPixel rings.
Or is that incorrect.
I snagged your Blinka event loop image. I link to the overview page twice. but I wanted to include a basic explanation on this page so folks aren't necessarily required to refer to another page to have a basic understanding of what's going on.
tulip sleet
<@!329766224093249548> Is it ok to use the word "simultaneous" when referring to...
the animations look simultaneous, to the user, but the two pieces of code are not running simultaneously: they are taking turns. It's a fine point, but one I emphasize several times.
idle owl
the animations look simultaneous, to the user, but the two pieces of code are no...
That is why I asked.
I will remove it.
@tulip sleet I made more changes to the code, but not functional changes, simply variable name changes. I tested whether GitHub can host GIFs (it can), and one step further, you can use the "remote URL" to load the GIF directly from GitHub in Learn. So I can make two GIFs, one for an example of a built in button, and one for an example of an external button, and then have the template-filler choose which one matches the wiring diagram. 🤯 Anyway, point is, I'll submit the GIFs with the code update as a single PR to Learn. I'll tag you on it for the code review since you got tagged by Limor on the last tweak I made. Should be sometime Monday or Tuesday if meetings go long or whatever. FYI.
idle owl
Also, once I'm done with this template page, I'll have you review it for content accuracy before handing it over the Anne and Limor.
tulip sleet
neato on the GIF!
idle owl
Right? Didn't think that whole process would work, but surprise, it worked.
lone axle
Has anyone seen this error attempting to build circuitpython from main? I'm sure something in my environment is messed up or out of date, but not sure what.
Traceback (most recent call last):
File "../../tools/gen_nvm_devices.py", line 2, in <module>
import cascadetoml
File "/home/timc/.local/lib/python3.8/site-packages/cascadetoml.py", line 12, in <module>
import typer
File "/home/timc/.local/lib/python3.8/site-packages/typer/__init__.py", line 12, in <module>
from click.termui import get_terminal_size as get_terminal_size
ImportError: cannot import name 'get_terminal_size' from 'click.termui' (/home/timc/.local/lib/python3.8/site-packages/click/termui.py)
make: *** [../../supervisor/supervisor.mk:75: build-pyportal_titano/genhdr/devices.h] Error 1
make: *** Waiting for unfinished jobs....
there is an issue with the exact versions of click and typer used . See the requirements-*.txt files for what is required now. But the typer person fixed the problem, so if you upgrade typer, it should be ok.
so we can undo the pins. See https://github.com/adafruit/circuitpython/pull/6210
@onyx hinge typer is fixed: https://github.com/tiangolo/typer/pull/375
lone axle
Installing click 8.0.4 as noted in requirements seems to have gotten me past that issue. Thank you!
tulip sleet
click made an API change without incrementing the major version number (which we didn't pin anyway, but still...)
Click 8.1.0 should work with typer 0.4.1
idle owl
@blissful pollen Is there another word for "returns" in the last sentence here? "await.gather(...) runs until all the listed tasks have finished. If gather returns that means all the tasks listed have finished." I basically get what it means, but I think it might confuse newbies.
"stops" isn't right, because it returning doesn't mean it's halting.
Neither is "finishes" or any other similar word, I think.
"restarts"? Or is that sometimes wrong.
blissful pollen
<@!228322543507865601> Is there another word for "returns" in the last sentence ...
A function does normally return but I know what you mean about maybe being confusing. Returns from running or something? Or maybe completes?
It would never restart though (unless it was in its own loop)
blissful pollen
No problem
onyx hinge
<@!426200547024961539> typer is fixed: <https://github.com/tiangolo/typer/pull/3...
yay! I saw that and meant to PR the reversion of our workarounds, but didn't get to it yet. (I didn't get to much today, somehow)
jaunty juniper
should the creation of special files on the drive be conditioned by CIRCUITPY_USB or whatever it's called ? You know, .Trashes and stuff ?
orchid basinBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
onyx hinge
Any bright ideas what to do about board.DISPLAY? this has vexed me as well. Changing the type of an object to NoneType is not quite the same as changing the object to be None, as None is also supposed to be a singleton.
Adding a generic "DeinitedObject" type could be interesting. Virtually any operation on it would raise a DeinitedError. Could save code, by removing the need to check for deinit explicitly in any functions.
stuck elbow
should have been a callable to begin with, like I2C and SPI
manic glacierBOT
onyx hinge
Adding a generic "DeinitedObject" type could be interesting. Virtually any opera...
barely a size advantage, after converting busio (+12 bytes free flash on trinket m0)
low sentinel
@lone axle sorry for sending you on a goose chase! I didn’t know you were doing a live stream, would have joined to chat & help clarify any curiosity. Looks like my comments cost you time.
(e.g., I think I meant “also” rather than “instead” :-p it’s just in my brain somehow more important in refresh areas so I wrote the wrong word)
lone axle
<@!382939733107408897> sorry for sending you on a goose chase! I didn’t know you...
No worries. I appreciate you're help!
I don't always know what I will work on ahead of time. I had a few ideas going into friday and ended up on that one. But if you are around one week it would be great to have you in the chat.
I'm definitely still "fumbling through" core stuff a bit so it's bound to take me some time one way or another, but I feel like I get better understanding each time I dive in to something like this so moving in the right direction 😄
@low sentinel do you think I have the right general idea about this being something that could be handled inside of Group rather than adding the hidden property to the vectorio shapes? It feels to me like Group should be able to manage that for it's children without them needing the flag specifically. Though the flag on the child elements can be helpful as well so maybe it's worth having it even if not strictly necessary.
manic glacierBOT
stoic rain
I'm looking at trying to get the WiFi/BLE functions of the Seeed WIO Terminal working with CP. Based on what I've found in digging into this, it should be possible, since the SAMD51 in the WIO seems to talk to the RTL8720D via an embedded remote procedure call API over UART (both UART and SPI are connected, but I think the latter is just for flashing the firmware).
Apparently, the more recent RTL8720D firmware supports both WiFi and BLE at the some time. Looking at the PyPortal as an example, it seems like the best way to approach this would be to implement a new class in adafruit_airlift to provide an implementation for adafruit_ble, and then maybe create something akin to adafruit_esp32api (adafruit_rtl8720api?) for the WiFi side of things.
Not sure I can even make this work yet (I need to dig into that eRPC code a bit more), but would love thoughts on how to best pattern this.
proven garnet
@idle owl, @trim elm - wanted to get the rest of the patch stuff done in time for Monday, in case you want to go over that as well:
Proposed changes to cookiecutter:
https://github.com/adafruit/Adafruit_CircuitPython_TSC2007/pull/2
Proposed changes to existing libraries (patch):
https://github.com/adafruit/Adafruit_CircuitPython_BME680/pull/48
lone axle
I'm looking at trying to get the WiFi/BLE functions of the Seeed WIO Terminal wo...
I think you have the right idea. about needing to implement something like adafruit_esp32api and adafruit_rtl8720api is probably what that library would get named if it were developed by adafruit folks, but I would guess that is unlikely unless they end up manufacturing a device with that type of wifi chip on it, or some member of the community takes particular interest in it. It could have a different name if developed by the community as well.
As far as best pattern I would probably lean toward trying to mimic the API in the esp32spi library if possible, but it could stray from that API if there are technical differences that necessitate it.
onyx hinge
what a weird bug, dan
tulip sleet
what a weird bug, dan
very confusing until I realized what was going on... tomorrow I will spend some time on it
it hangs when I gc, when the pins are connected like that. totally weird
onyx hinge
that doesn't make much sense
tulip sleet
it didn't hang when I picked a different pin. I thought it was a GPIO0 issue, but the different pin was not connected, so it worked
onyx hinge
so if it took you a bit to realize the pattern that's not surprising
tulip sleet
and it's the TX pin. I would have thought maybe it was due to noise on the RX pin causing a lot of interrupts
or maybe the TX pin transitions are bleeding into the RX pin
onyx hinge
some microcontrollers can back-power through protection diodes, maybe it draws a bunch of current?
tulip sleet
maybe, maybe, but why only on gc?
onyx hinge
oh yeah, a tx looped back to its own rx could be interesting. you could also try a tx-only or rx-only uart
or an arduino program
tulip sleet
good idea, first thing is that I will see if it's looping or something. Just got the SWD connector working. I had another Feather with an SWD box but I coudn't find it
manic glacierBOT
manic glacierBOT
manic glacierBOT
@anecdata @dhalbert - Will go ahead and close this issue.
The crashes do appear to be duplicates of the issues mentioned. As I said in previous post will sync up and try with latest sources.
however it is still very unclear on what is the proper way to use two hardware uarts of the RPI2040?
From the pinout page, I believe I have options on which pins to use for each:
...
idle owl
<@!330227457296957440>, <@!362023826105827340> - wanted to get the rest of the p...
Thank you! I'll add this to In the Weeds as well. Will you be able to voice chat in the meeting today, at least for in the weeds? Or is it too late in the day for you to be available other than listening in?
tulip sleet
Reminder that we are having a CircuitPython Weekly meeting today, in about two hours, 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/1em2Z3EF5T_eaz04e-XdesuIZrxCEB_3plFVzBiat_jc/edit?usp=sharing
idle owl
TFW you're typing in the notes doc and then the host adds all the things. My typing is many pages away now. 😄
tulip sleet
maybe this is a new form of teleportation
@onyx hinge I see all your PR's, getting ready for the meeting, so I probably won't review them until after
onyx hinge
@tulip sleet thanks, no hurry!
looks like 7.2.x should be merged to main after those go in as well.
low sentinel
@lone axle Yeah, I was hoping to keep hidden out of the vectorio APIs and let it be handled by Group from the Python point of view.
Managing dirty areas at Group makes sense imo; ideally Group could handle it but you might have some challenges, e.g., with cleaning up previously drawn areas especially if there’s movement. As a fall-back, it seems reasonable to put a c-level boolean in the vectorshape struct. Would prefer avoiding it BUT hidden/not-hidden from group is an important enough feature to pay the memory for it 👍
onyx hinge
huh I wonder why some of those builds are failing. the error seems to be FileExistsError: [Errno 17] File exists: '/opt/hostedtoolcache/Python/3.10.4/x64/bin/python' -> '/home/runner/work/circuitpython/circuitpython/.idf_tools/python_env/idf4.4_py3.10_env/bin/python' from a virtualenv creation attempt. https://github.com/adafruit/circuitpython/runs/5818926795?check_suite_focus=true
orchid basinBOT
lone axle
<@!382939733107408897> Yeah, I was hoping to keep `hidden` out of the vectorio A...
okay thank you! appreciate your perspective. I'll make some more attempts at doing it from the Group. I feel like I was close with that approach, just need a tweak somewhere to make it realize it needs to redraw. Will fall back to property on vectorshape mimicing the ones from TileGrid and Group if I don't end up having success doing it from inside of Group only.
idle owl
@tulip sleet I took the updated libs out of the notes doc. Over a page is too many, nearly 3 pages is untenable. I deleted them from the notes doc, and added a link to the library report on S3 if folks are interested in seeing the list.
tulip sleet
<@!329766224093249548> I took the updated libs out of the notes doc. Over a page...
That is good. I was thinking of that, but your decision is more authoritative
idle owl
No worries. Always happy to help.
proven garnet
I could be in the voice channel :)
idle owl
I could be in the voice channel :)
Keen! That might be helpful for discussing In the Weeds. I was pretty thorough, but I didn't do the work. 🙂
stoic rain
I think you have the right idea. about needing to implement something like `adaf...
Cool. I'm still digging into the eRPC code for the WIO Terminal's Arduino core, so it'll be a little while before I can say for sure, but I think keeping close to the esp32spi lib will work. Seeed implemented a library that mimics the ESP32 API (rpcWiFIi) so I might take a look at that, too.
digital shoreBOT
mental nexus
Hi all!
turbid radish
Please subscribe at adafruitdaily.com
idle owl
It's not available everywhere yet. FYI.
onyx hinge
you pasted exactly the information that was needed for me to guess at the problem so yay
errant grail
Algorithmic Problem Solving with Python https://eecs.wsu.edu/~schneidj/swan/index.php
onyx hinge
oh! that's much more my speed 😁
idle owl
you pasted exactly the information that was needed for me to guess at the proble...
Pasting didn't help either of us figure it out, so still... good on you!
onyx hinge
88 boards = 1 piano
gilded cradle
🙂
gilded cradle
oof a $140 textbook !? https://www.pearson.com/us/higher-education/product/Schne...
You can get it a bit cheaper if you shop around: https://www.amazon.com/Introduction-Programming-Using-Python/dp/0134058224
onyx hinge
idle owl
@lone axle I can point you to where you need to look, I believe. Not sure I can help, but I can at least get you started in the right direction.
lone axle
<@!382939733107408897> I can point you to where you need to look, I believe. Not...
awesome! Definitely appreciate a point int he right direction.
idle owl
If you didn't already though, you should start by seeing if the API can even return that information.
errant grail
Happy yearly birth date @onyx hinge !
idle owl
I assume the API can do that, since it has a ton of info, but assuming has wasted a lot of time in general for me...
onyx hinge
I haven't managed to pick up an esp32c3 yet so I haven't tested.
lone axle
If you didn't already though, you should start by seeing if the API can even ret...
Good point, I will look into the github API before I start digging back into adabot. Would be good to verify that data is available before spend too much time figuring out where / how to add it to the page.
idle owl
Good point, I will look into the github API before I start digging back into ada...
Exactly 🙂
onyx hinge
absolutely right to file issues
tulip sleet
Good point, I will look into the github API before I start digging back into ada...
also if you can find out a milestone for a PR, that would be a good label too (e.g. the 8.0.0 PR's that are in abeyance for a while).
lone axle
That would be great. I'll check into milestone data availability as well.
onyx hinge
nice!
idle owl
Congrats @proven garnet!
ember iris
Congrats! That's super exciting!
lone axle
Preliminary it's looking promising for both draft and milestone data. I figured out how to access both with the gh cmd line utility. I think that utilizes the Github API "under the hood" to get it's information.
idle owl
Preliminary it's looking promising for both draft and milestone data. I figured...
Excellent.
lone axle
I do like the idea of using a minimal .gitignore in the repos as well.
I didn't know about that personal ignore file until Jeff mentioned it on a PR at one point.
jaunty juniper
and there's .git/info/exclude
ember iris
Are there any comments as to why there are added blocks in the .gitignore? Some of my personal ones have comments address what a block is addressed for.
It doesn't solve the problem one way or the other, but it might make it easier to see why I have something in a gitignore.
lone axle
Is it possible to make a CI step that validates that there are no changes inside of .gitignore? Or perhaps even more broadly we could have a larger list of "infrastructure" files that are all checked during CI and get flagged if they are chanaged.
tulip sleet
we can add a comment to cookie-cutter .gitignore that explains how to add things to your personal .gitignore
onyx hinge
https://git-scm.com/docs/gitignore
Patterns which a user wants Git to ignore in all situations (e.g., backup or temporary files generated by the user’s editor of choice) generally go into a file specified by core.excludesFile in the user’s ~/.gitconfig. Its default value is $XDG_CONFIG_HOME/git/ignore. If $XDG_CONFIG_HOME is either not set or empty, $HOME/.config/git/ignore is used instead.
lone axle
a comment in the file that points people toward how they can set up their personal .gitignore sounds like a helpful thing to me as well.
gilded cradle
Thanks
onyx hinge
I'm out!
proven garnet
Thank you!
gilded cradle
Me too
ember iris
Thanks all!
idle owl
So major IDEs that folks might use are PyCharm, VSCode, Mu, vim, Atom... what else? So far only the first two add spurious files when using them. (At least with my minimal usage. Am I missing something about the others?)
Nano didn't leave anything either.
What do other folks use?
ember iris
Sublime, Jupyter notebooks but I doubt they'd be used here, Emacs, Spyder. (Based on python developer survey, excluding what you've already covered)
ember iris
Googling sublime cache files gitignore show's a good number of example gitignores: https://github.com/github/gitignore/blob/main/Global/SublimeText.gitignore which might help with figuring out what ide's generate what cache's
idle owl
Hmm. Sublime isn't leaving any files when I'm editing here.
Oof that's a lot of files.
That the gitignore is saying it might leave.
jaunty juniper
yeah it's probably special cases ?
ember iris
Honestly I wouldn't worry about adding them
idle owl
I'm not getting any files with simple changes.
jaunty juniper
you might need to setup the dir as the root of a project to get them
idle owl
from Sublime. So... probably not worth adding.
ember iris
I wonder if that's if you try to git add * in the same folder as the ide
idle owl
you might need to setup the dir as the root of a project to get them
Oh, so actually use it like it's meant to be used.
Fair enough.
lone axle
So major IDEs that folks might use are PyCharm, VSCode, Mu, vim, Atom... what el...
I see a fair amount of people mentioning using Thonny. Don't know if it creates any files though.
idle owl
Here's my thought on Sublime. It's not currently in any of the libraries.
So I'm not adding it here.
ember iris
I think that's reasonable
idle owl
I see a fair amount of people mentioning using Thonny. Don't know if it creates ...
Oh right, Thonny. I have it on this machine, I can check.
jaunty juniper
agreed
idle owl
Don't see anything from Thonny either. When I use it, or in the current pile of gitignore entries.
ember iris
Plus if it does generate some cache file, you can also just not add it to a commit. Foamyguy's comment about updating the CI so it fails if there's a change to the gitignore would help with this as well once the minimal gitignore is setup.
idle owl
Agreed.
idle owl
How likely is one to run into .pyc files when working with the libraries? I looked up what it is, and I don't quite follow how one might end up with one in a library repo.
Same with __pycache__ ?
proven garnet
How likely is one to run into .pyc files when working with the libraries? I look...
Running code using the library. I think it's usually in a Gitignore because there's never a reason someone should need to track compiled python
idle owl
Hmm.
tulip sleet
Here is the notes document for next Monday’s CircuitPython Weekly meeting. It is at the normal time of 11am Pacific / 2pm Eastern here on Discord. Everyone is encouraged to attend! Please add your hug reports and status updates even if you’ll be attending the meeting - it’s super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and we’ll read them off during the meeting. Hope to see you there! <@&356864093652516868>
https://docs.google.com/document/d/1m1PCDAFk9kLs3MF1vEx9e0UW1WTURFB-eTORSQe8a7o/edit?usp=sharing
proven garnet
But I guess if the target is an MC it might be different. Sphinx may generate it though?
idle owl
But I guess if the target is an MC it might be different. Sphinx may generate i...
Sphinx does not.
tulip sleet
I get those if I am testing the library with Blinka, etc.
ember iris
I think both extensions are worth keeping in the gitignore since so many scripts can be run or generated from cpython
idle owl
I'll include them then.
idle owl
Last few files I think aren't needed anymore because they're PyPI related, and that's handled by the CI now. You don't do that locally.
Though I guess Scott was doing the one that used pyproject.toml manually or something. We shouldn't be doing it locally anyway.
What would a bundles file be?
ember iris
looks like it might be something to do with javascript?
idle owl
What are we doing with JS?
Pretty certain this is all manual-PyPI related: #build*/ #dist #.eggs #**/*.egg-info
We were dealing with those files when we first put all the libs up on PyPI.
We should no longer be dealing with them.
tulip sleet
i think if you pip -e install ., some of that might show up. Need to test
idle owl
Last questionable inclusion is .env. I suppose that one should be up to folks to do globally?
tulip sleet
I'll try right now
idle owl
Ah, I was going to say that 😄
I'll let you test.
Instructions for creating a .gitignore_global - beginning or end of this .gitignore file? I feel like beginning since it's not supposed to be edited, so we're kinda of saying "Before you even look at this file, before you even think about editing it... maybe you need to follow these steps."
tulip sleet
*.egg-info shows up when i install the BLE library
tulip sleet
I think so, if it supports **
idle owl
Dunno why it's got extra ** that the others don't...
I'll take it out.
Nothing else uses it, but it's obviously supported.
Conclusions on bundles?
ember iris
What are we doing with JS?
https://github.com/adafruit/Adafruit_CircuitPython_APDS9960
https://github.com/adafruit/Adafruit_CircuitPython_Bundle
https://github.com/adafruit/CircuitPython_Community_Bundle
Those are the libraries where I found bundles in the gitignore.
I *think* that the APDS9960 doesn't need it, I don't know about the libraries bundles, and I think those are out of scope here
idle owl
This file is so short. Heh. I mean I added comment headers for everything in there, and am about to add how to generate a global gitignore, but still. I can't believe how bloated it got across the libs over time.
Hmm. Yeah, the patcher doesn't touch the bundles, and they could be different, though they might want to be edited as well, while we're in this.
Windows and Linux don't add any stupid files, right? Only MacOS does that?
ember iris
I would think it's safe to not include bundles then. I don't know js well enough, but I can't see it being needed in the cookiecutter-generated gitignore
idle owl
This will be the cookiecutter-generated gitignore. Not the actual cookiecutter one.
ember iris
Right, that nuance get's confusing. I still think that holds
idle owl
🙂 I can see that.
ember iris
it's only a monday and my mind is melty. It's going to be an interesting week ahead of me.
I don't know about windows vs linux, I don't notice weird temp files being generated on linux though (but I'm unobservant so someone else should weigh in)
idle owl
I think it's only Mac.
Explanation make sense? ```# If you find that there are files being generated on your machine that should not be included in
your git commit, you should create a .gitignore_global file on your computer to include the
files created by your personal setup.
First, create a .gitignore_global file somewhere convenient for you, and add the rules for the files you
want to exclude from git commits.
Second, congfigure Git to use the exclude file for all Git repositories by running the
following via commandline, replacing "path_to_your" with the actual path to your newly created
.gitignore_global file:
git config --global core.excludesfile path_to_your/.gitignore_global```
jaunty juniper
Last questionable inclusion is `.env`. I suppose that one should be up to folks ...
every library includes that instruction in the "building locally" section:
python3 -m venv .env
so I think .env should be in it since we make people create it
idle owl
every library includes that instruction in the "building locally" section:
```sh...
Oooooh you are correct.
Agreed entirely.
I had forgotten about that bit.
idle owl
so I think `.env` should be in it since we make people create it
Thanks, Neradoc.
tulip sleet
https://docs.github.com/en/get-started/getting-started-with-git/ignoring-files has pointers to some common lists, including https://github.com/github/gitignore and https://gitignore.io (!).
idle owl
Yeah. I have that page open.
tulip sleet
note that the global gitignore file does not have to be called .gitignore_global.
idle owl
These instructions need to be short. Should I include that?
It can be called whatever you want it to, as long as you change the command to match.
But that bloats the instructions a bit.
tulip sleet
"called, say, .gitignore_global." Windows used to have trouble with files with leading dots, I'm not sure if it still does
idle owl
Oof. Ok. I'll have to specify the command needs to match the filename as well.
tulip sleet
looks like it is not a problem anymore. OK, never mind
idle owl
Oh. Oki.
tulip sleet
I used to have to do _emacs instead of .emacs, but that is from years ago.
idle owl
Fair enough.
tulip sleet
out for food shopping
idle owl
Oof, that gitignore.io site produces HUGE files, like instead of .idea, it includes literally everything possibly contained in the .idea directory listed out separately.
Apparently build/ is a Python thing.
tulip sleet
gitignore.io did not look that helpful to me either
idle owl
But the Python list included eggs which, afaik, is a PyPI thing. So who knows.
idle owl
Explanation make sense? ```# If you find that there are files being generated on...
Any other comments on these instructions?
tulip sleet
"congfigure"
idle owl
Thank you. It's not showing misspelled words though it usually does.
I think it doesn't really know what to do with this file. other than comment highlighting.
tulip sleet
I would be prescriptive, say: "Do not include files and directories created by your personal work environment, such as the IDE you use, except for those already listed here."
something like that
idle owl
Ok. That sounds good.
tulip sleet
ok now really going
idle owl
Is it something weird about cookiecutter that makes it not show me a diff on the modified .gitignore file within the repo generation code?
git diff returns nothing. I tried from both the main dir, and the same dir as the file.
OK, weird, now it's returning updates made to the file since git add. Bleh. Shrug.
Oh, now it's working. Must have been something to do with adding it before diffing it.
¯_(ツ)_/¯
jaunty juniper
Oh, now it's working. Must have been something to do with adding it before diffi...
maybe you wanted git diff --cached ?
idle owl
maybe you wanted `git diff --cached` ?
It's possible.
jaunty juniper
otherwise diff diffs between the staged (added) files and the current files
idle owl
good to know.
Was going to ask for another look at the .gitignore file, but I suppose the PR is as good a place for that as any.
For anyone who is interested: https://github.com/adafruit/cookiecutter-adafruit-circuitpython/pull/179
gloomy shuttle
Is there a process or best practice for taking the changes made to the Adafruit libraries and applying them to community libraries like these changes to .gitignore?
idle owl
Is there a process or best practice for taking the changes made to the Adafruit ...
Not at the moment! If we think it should be the same across community libraries as well, this one would be an easier one to do since it's replacing an entire file, vs changing parts of an existing file. But our Adabot patch system doesn't handle the Community bundle, unfortunately. Which is what we use to update the Adafruit libs.
Once it's merged to cookiecutter, we can discuss applying it to the Community Bundle if everyone agrees. Or we can have maintainers update it as desired.
gloomy shuttle
I knew adabot doesn't handle them.
idle owl
That's part of it, is that they're all maintained by individuals, who may not want the update, necessarily. Cookiecutter will generate it for new libs though.
gloomy shuttle
I know there are other changes that I haven't kept up on like changes to CI (black/click version, etc...)
idle owl
Understood.
This is definitely something worth discussing with the group. We could try to come up with something to at least notify folks maybe? I don't know how difficult that would be though.
orchid basinBOT
gloomy shuttle
Thanks and I agree. I was just making sure there wasn't something in place that I had missed
idle owl
Kind of like Dependabot, but not so integrated.
Ah fair enough! You're not missing anything.
shut dew
@idle owl just to chime in, git core.excludesfile has a default location $XDG_CONFIG_HOME/git/ignore aka ~/.config/git/ignore. Folks that use that location don't need to configure it explicitly at all
idle owl
<@!330227457296957440> just to chime in, git core.excludesfile has a default loc...
Thanks for the input! I'll think on whether it makes sense to include that or not. I got those instructions off of the Git documentation site, though I'm pretty sure they cover that option as well further down the page. I'll see if I can include it and keep it concise. Thank you again!
Make that from the GitHub documentation.
And I'm wrong, they don't cover this.
shut dew
yeah, the GitHub documentation seems to be old or conservative as it doesn't mention $XDG_CONFIG_HOME at all, but that makes sense as it's relatively recent
shut dew
some tool/apps are more or less aggressive about migrating to it.
idle owl
Ah it's a newer thing? (I guess you essentially said that.)
shut dew
'relatively' i think the XDG standard has been around for at least 10 years now, but lots of software is much older and slower moving
stuck elbow
it's ~/.git
idle owl
I don't have that either.
idle owl
'relatively' i think the XDG standard has been around for at least 10 years now,...
I thought that was a variable to be replaced by the user with their home directory. I didn't realise it was an actual thing.
shut dew
it's a variable in that it can be redefined where the user wishes, but ~/.config is the default value if undefined
idle owl
Ahh. Ok.
shut dew
an overview of the standard, and which apps are compliant with it https://wiki.archlinux.org/title/XDG_Base_Directory
idle owl
Thanks!
manic glacierBOT
Got to Blynka!
The REPL and Blynka logo are working. The image is offset due to some timing weirdness that shifts the screen whenever CircuitPython does some processing, or if you type in the REPL (see previous update).
Todo: When the code finishes, the display goes weird, doesn't update anymore and doesn't show the REPL.
orchid basinBOT
thorny jay
Card library for Circuit Python
tulip sleet
@onyx hinge I labelled four PR's with 7.2.x milestones. I think they are all we want for bug fixes for 7.2.5, so once they are approved and merged, I can make another release, and you can finish the 7.2.x -> main build. I could make another beta at that time.
onyx hinge
<@!426200547024961539> I labelled four PR's with 7.2.x milestones. I think they ...
Good morning. I'll look
onyx hinge
@tulip sleet I can mash 'merge' on those PRs if you like, they were very simple and to the point.
tulip sleet
<@!329766224093249548> I can mash 'merge' on those PRs if you like, they were ve...
Sure, thanks! go ahead, I did test them both. There is a big backlog of builds now
tulip sleet
I will be out for a while early afternoon, will try to do releases after that, after things have caught up.
onyx hinge
should we cancel some of the older 7.2.x branch builds?
tulip sleet
you mean the merge builds? We now have auto-cancel, which cancels and restarts if there is a new push.
tulip sleet
it's great! Thanks to microdev for that
tulip sleet
oh okay, good!
I thought you closed it by accident. No matter.
onyx hinge
I think I accidentally force-pushed the branch until it was even with master, which caused github to close it.
onyx hinge
ports/atmel-samd/mpconfigport.mk:ifeq ($(TRANSLATION), ko)
``` interesting, I just noticed that the code to exclude the fonts for these translations is port-specificthe "ja" translation has 454 distinct glyphs but tilegrids have only 256 distinct tiles so it can't be displayed even if the font fits in ROM
(well, for some specific build it has 454 glyphs, it probably varies according to features)
onyx hinge
@tulip sleet curious what the lock is guarding against here. ```void common_hal_keypad_keymatrix_reset(keypad_keymatrix_obj_t *self) {
const size_t key_count = common_hal_keypad_keymatrix_get_key_count(self);
supervisor_acquire_lock(&keypad_scanners_linked_list_lock);
memset(self->previously_pressed, false, key_count);
memset(self->currently_pressed, false, key_count);
supervisor_release_lock(&keypad_scanners_linked_list_lock);
doesn't scan happen via background task? or is it via irq/timer?
tulip sleet
it was against another keypad scanner being added to the list while a scan was in process. Maybe it is not necessary.
i might have written it thinking that something might be done at interrupt level
there were many versions of the code
or maybe removed from the list
onyx hinge
.reset() doesn't clear the event queue too?
orchid basinBOT
onyx hinge
weird, how is that happening? is pypi not serving up the same versions everywhere?!
wow the versions are WILDLY different. ```['adafruit-circuitpython-busdevice-5.1.1', 'awscli-1.22.38', 'botocore-1.23.38', 'charset-normalizer-2.0.9', 'cryptography-36.0.1', 'jmespath-0.10.0', 'pyOpenSSL-21.0.0', 'pyftdi-0.53.3', 's3transfer-0.5.0', 'tomli-2.0.0', 'typer-0.4.0', 'urllib3-1.26.7']
['adafruit-circuitpython-busdevice-5.1.8', 'awscli-1.22.88', 'botocore-1.24.33', 'charset-normalizer-2.0.12', 'cryptography-36.0.2', 'jmespath-1.0.0', 'pyOpenSSL-22.0.0', 'pyftdi-0.54.0', 's3transfer-0.5.2', 'tomli-2.0.1', 'typer-0.4.1', 'urllib3-1.26.9']
i am baffled. I thought it was GitHub at first, but it seems like pypi is going a bit crazy
onyx hinge
pyOpenSSL-22.0.0 was released January 29, so some database seems to be serving up months-old versions instead of new ones. does pypi have a status page?
https://status.python.org/ yeah something's going wrong
Monitoring - Applications are coming back online and we are monitoring for stability.
Apr 5, 15:41 UTC
Identified - Core failed service has been identified and is coming back online. Next we'll bring up the ancillary services. Once all our automation services are online we can begin to bring the applications back.
Apr 5, 15:38 UTC
Investigating - All backend services for PyPI are down due to a cascading failure in our deployment tooling. We are investigating and working on restoring service.
Apr 5, 15:25 UTC
tulip sleet
yeah bad news
I will do merges when things finish, but I won't start release builds until this is fixed
idle owl
Oof.
lone axle
@idle owl I have created a new (currently hidden) page in the sharing a library guide and compiled together all of the notes and other typing related things that I know about: https://learn.adafruit.com/creating-and-sharing-a-circuitpython-library/typing-information It's not quite done yet. The two things I know for sure that I still intend to add are an example of the typing import with the try/except that we use, and a section that explains Union, Optional and similar other typing helpers that get used sometimes.
When you have a moment if you'll take a look at what I have so far and let me know if you have any feedback.
There is likely some copy editing that could be done to improve things but I want to make sure everything so far is headed in the right direction and see if there is anything else specific you can think of that would be good to add to it.
idle owl
<@!330227457296957440> I have created a new (currently hidden) page in the shari...
Thanks so much! I'll take a look once I finish what I'm working on right now 🙂
lone axle
Actually a 3rd thing that I plan on adding comes to mind now as well: A section that talks about how to find the correct types. I will explain the process of looking in the code and examples or other usages for clues about what the proper type is and also show how to use type() in REPL or in modified library code to check on actual types being used.
idle owl
This all sounds great
onyx hinge
I suppose I should test this code I'm working on, next. simplifies and drops 600 bytes of code (in a non-LTO build) from keypad. but of course maybe it's so much smaller because I deleted something and didn't put it back 🙂
idle owl
@tulip sleet When you're available, I'd appreciate a final review of this code: https://github.com/adafruit/Adafruit_Learning_System_Guides/pull/2135 (No functional changes, only variable name changes.)
idle owl
Once that PR is merged, the next step is to add it to a board guide and fill it in. I'll have you review it for accuracy before Anne and Limor.
idle owl
@lone axle If one wanted to load a Community library onto a Pi for use with Blinka, how would one go about doing that? Do a pip from within the library directory or something? I have someone on the forums asking about it. If you know, do you mind replying?
I at least assume that's what this person is trying to do.
I doubt they're using bare metal.
Mostly because most folks are using Blinka.
lone axle
<@!382939733107408897> If one wanted to load a Community library onto a Pi for u...
There are a few possible ways, but pip install is probably the easiest if they already have pip set up on their pi (which they should if the followed the setup guide for Blinka/Pi)
and yep link me and I can respond.
idle owl
They do apparently.
Thanks so much! @lone axle https://forums.adafruit.com/viewtopic.php?f=60&t=189997
jaunty juniper
you have to look at the library to see if it is on pip to begin with
idle owl
you have to look at the library to see if it is on pip to begin with
They tried pip, couldn't find it.
idle owl
@onyx hinge As everyone else is gone or idle, I'm pinging you with this one. nRF board question. On reset, begins by flashing yellow to allow for entering safe mode, then blue to allow for entering pairing mode, before fully booting into CircuitPython. Press reset during blue, what should be happening here? The LEDs flash again, but there's a moment of solid blue that time, and then in a VERY short period of time, it goes back to running CircuitPython code. That doesn't seem right. Updated to 7.2.4, and tried on CPB and CLUE, same behavior.
I feel like I remember pairing mode lasting a lot longer than that. As in... long enough to actually pair and remain paired.
jaunty juniper
I feel like I remember pairing mode lasting a lot longer than that. As in... lon...
it's on for good once you reboot in pairing mode
idle owl
No... it's not. Not right now, anyway.
jaunty juniper
well it should be
idle owl
It almost immediately bounces back to running code.
jaunty juniper
yes but the pairing is on
idle owl
Ok.... I'll try again.
onyx hinge
I'm glad neradoc is answering because I don't know
jaunty juniper
or it should be
idle owl
Ok, it paired properly that time. I think.
The app isn't seeing it now, but that's a separate issue.
OK, thanks @jaunty juniper. I guess it was a me issue. That's what I get for not assuming that this time, like I usually do.
I guess we need to document that somewhere unless we did and I missed it. Which is probably the case. Because that is confusing.
To me anyway.
jaunty juniper
it might be flaky, it's bluetooth after all 🤔
idle owl
Also valid.
I guess I was expecting it to keep flashing blue or be solid blue until it paired. That was my mistake.
onyx hinge
the LED codes probably are all different since the last time you did it, and guess who gets to make sure the guides are clear !?
if you had a mirror you'd be looking at her
jaunty juniper
it might be clearer if booting in pairing mode made it wait until you connect to it with an app, while (!Serial) style
idle owl
the LED codes probably are all different since the last time you did it, and gue...
Also doesn't help if I can't figure out what is actually happening. Can't update a guide without the right info.
idle owl
it might be clearer if booting in pairing mode made it wait until you connect to...
I actually know what you're referring to!
jaunty juniper
like doing the reset dance is an intentional action at a point in time, you then know to connect to it, while putting it in pairing mode is a mode that is otherwise invisible afterwards (outside of scanning with the app when BLE is in a good mood)
idle owl
I'm certain Scott has Opinions about this, and its current behavior is probably what he wants.
onyx hinge
ValueError: IO44 in use
Hmmm I wonder why this happens. Maybe it's because I made a debug build, and this is one of those 'shared with jtag' pins? That's my current theory.
rebuilds circuitpython
yup that was it
idle owl
@lone axle Wow this page is beefy! The only suggestions I have are nitpicky formatting things. I think when referring to Python, it should be capitalised at all times. As well, I think it's ok in this case to use the Heading 1 for the ones you have a Heading 2 on currently, and Heading 2 for all the current Heading 3s. That will make all headings linkable, which I think might come in handy. Otherwise, other than me not knowing accuracy because I don't know typing information all that well, it looks great to me!
lone axle
<@!382939733107408897> Wow this page is beefy! The only suggestions I have are n...
Thank you! Appreciate the feedback I'll incorporate that stuff when I get back to it.
idle owl
Sounds good!
tulip sleet
@idle owl @jaunty juniper The point of the blue delay was not to enter pairing mode, it was to erase the pairing info. You can also do that by _bleio.adapter.erase_bonding(). I will double-check to see if anything has changed about this. It is explained in this diagram, from here: https://learn.adafruit.com/welcome-to-circuitpython/troubleshooting#circuitpython-7-dot-0-0-and-later-2978455-21
idle owl
<@330227457296957440> <@199811382793863168> The point of the blue delay was not ...
"Pressing reset during the BLUE blinks will clear Bluetooth information and start the device in discoverable mode" so kinda both?
This is for use with PyLeap, no one is going to add _bleio code to make it work with PyLeap. Defeats the purpose.
tulip sleet
that is the most correct, the lack of bonding info makes the host say, "do you want to connet with this"
idle owl
Hmm ok.
jaunty juniper
oh wait, you can connect at any time with the App once it's bonded, even if you power cycle it ?
tulip sleet
anyway, sounds like timing is messed up
some connections require bonding, like HID mouse and keyboard
i have not used PyLeap enough to know, but I think so too, so you can use it in a classroom
idle owl
I've been resetting it with the button, and it remains known to PyLeap, but that's not power cycling.
It pops up a "Do you want to connect to CIRCUITPYfoo1" that you have to say yes to when first using it
idle owl
anyway, sounds like timing is messed up
I'm not sure it is. I think it might be ok. I think the whole thing might have been a me-issue. Not understanding what was going on.
But maybe I'm now simply used to what it actually is doing, so I don't take issue with the timing. Because I had to make it work.
jaunty juniper
so blue reset makes it discoverable, which is required to pair it with pyleap, but it also erases all bonding, including bonding made through python ?
tulip sleet
yes, it erases all bonding info
idle owl
Which is awkward with PyLeap, because there's still an issue where you have to manually go in and "Forget" the device if you have to repair it at any time, or it won't connect. They're working on trying to fix it, but it might be an iOS thing.
tulip sleet
this is in supervisor/shared/bluetooth/bluetooth.c
idle owl
I'm not suggesting changing the behavior, simply relaying semirelevant info.
tulip sleet
there is a yellow blink for safe mode and a faster blue blink for "clear bonding and get into discovery mode". After you click reset during blue, the next time around (because it's a reset), I think it does not wait and show 1sec of blue
idle owl
It shows some period of solid blue before code runs.
tulip sleet
the first time or after pressing reset?
idle owl
Following pressing reset during flashing blue.
tulip sleet
yes, it takes a bit of time to do the discovery
idle owl
So flash blue, hit reset, starts up, some period of solid blue after whatever, and then code runs
But if you're not looking for it, it's easy to miss, I think.
tulip sleet
yes, that's what I see, but only after reading the code. Perhaps there needs to be a fuller explanation somewhere
idle owl
Probably something like that.
I'm not sure how things like this will be explained in PyLeap.
There's a mini tutorial if it can't find your board, which could include this.
They're updating it soon. So I will make a suggestion now.
Once I find out where to leave suggestions. Since they're not bugs. 😄
onyx hinge
https://docs.micropython.org/en/latest/library/micropython.html#micropython.opt_level do we specify any optimization level with mpy-cross? I was looking into whether it is possible to have mpy-cross eliminate if typing.TYPE_CHECKING: ... blocks but that involves changing the compiler so it can (A) assume TYPE_CHECKING is false and (B) enabling at least optimization level 1, which also deletes asserts. Going up to optimization level 3, which removes line number information, could be a memory savings to consider taking on samd21 builds with frozen files.
(part "A" is a big TODO too)
jaunty juniper
https://docs.micropython.org/en/latest/library/micropython.html#micropython.opt_...
does that include adding a typing library ?
onyx hinge
there's a "postmortem" of the pypi outage. https://status.python.org/incidents/mxgkk3xxr9v7
@jaunty juniper yes it adds a typing module with the only member TYPE_CHECKING and the value False
jaunty juniper
that reminds me, is supporting blinka on Micropython a goal ?
onyx hinge
I was under the impression that it worked, though I don't know to what extent.
jaunty juniper
like, currently when I tried to install the files on a pico, there was not enough free space on the drive
tulip sleet
@onyx hinge I am confused about your TYPE_CHECKING PR, could I ask you some q's?
onyx hinge
@tulip sleet it can wait, we're going to start making dinner in a second
tulip sleet
sure
onyx hinge
I don't blame you for being confused, my ideas come out of left field sometimes
tulip sleet
the original goal of Blinka was to support MicroPython! (if I remember right). That quickly morphed into SBC's, and MicroPython fell by the wayside.
It's still a goal, but a minor one
jaunty juniper
and for example are we actually using __future__ now in libraries, which is not in MP ? and typing is not on MP either
tulip sleet
from __future__ import typing is not a thing in CPython. I am not sure we should do it this way.
We want the code to be compatible with CPython
I added from __future__ import annotations because that is a thing
jaunty juniper
for the drive space issue, I was wondering if yould be relevant to some CI make a micropython release of blinka, with most boards definition files pruned (or if it could be installed like that with upip)
yes I mean from __future__ import annotations, it's not in MP
tulip sleet
(but the new PR is from __future__ import typing)
jaunty juniper
from typing import TYPE_CHECKING
tulip sleet
yes, a pruned version of blinka makes sense, but we have not done such pruning yet. We talked about doing it for some frozen libraries (like the CP library), but didn't come up with a satisfactory solution yet
onyx hinge
sorry for making you waste brain cycles on that Dan, except that obviously I wasn't going to come up with the try/except variation myself
tulip sleet
np, I get bad ideas when my blood sugar is low too 🙂
i think it was a good bad idea
the motivation is excellent
devout jolt
Hi! I have a question about pylint CI. I just submitted a PR on this library and it's complaining about Module 'struct' has no 'unpack' & Module 'struct' has no 'pack' on code I did not touch. https://github.com/adafruit/Adafruit_CircuitPython_BLE_Apple_Notification_Center
tulip sleet
Hi! I have a question about pylint CI. I just submitted a PR on this library a...
I don't see that; I just see that it failed black formatting. Have you been using pre-commit? https://learn.adafruit.com/creating-and-sharing-a-circuitpython-library/check-your-code
devout jolt
yep
I was hoping there was something wrong with my pylint setup and CI would be different
jaunty juniper
Hi! I have a question about pylint CI. I just submitted a PR on this library a...
I used to get that locally until I uninstalled an older version
tulip sleet
make sure black is up to date, should be 22.3.0 or so
devout jolt
I am at 22.3.0 yes
tulip sleet
I used to get that locally until I uninstalled an older version
older version of pylint? @devout jolt how about version of pylint?
devout jolt
but why is it failing in CI then too?
tulip sleet
because your pre-commit formatted it with the old black?
and that's what you comitted?
maybe pre-commit autoupdate ?
would fix things
devout jolt
I do not understand. shouldn't it be checking the files with its own version of black? And it's failing on code I did not touch. And it's saying "pack()" is not a member of "struct"
jaunty juniper
that (the pre-commit cache thing) was for your fail on struct, CI doesn't fail on struct
devout jolt
I cannot tell why CI failed, can you? https://github.com/adafruit/Adafruit_CircuitPython_BLE_Apple_Notification_Center/runs/5842541346?check_suite_focus=true
jaunty juniper
tulip sleet
devout jolt
yes what is the failure?
jaunty juniper
devout jolt
yes but why?
tulip sleet
because the files were not formatted to its liking
because your pre-commit is running an old version of black, per Neradoc
so you formatted it the old way, and it wants it the new way
devout jolt
sigh okay. lemme try again. Is it standard practice to always have to pre-commit autoupdate before pre-commit run --all-files to catch whatever the formatting-du-jour is? I don't have a good feel for the standard dev cycle here
tulip sleet
black hasn't changed its mind in a long time. This is new.
i agree with you that pre-commit could do a better job. They have some philosophical issue
about versions
devout jolt
I notice in the commit log for that library there were some black updates in the pre-commit-config.yaml but I assumed that would be taken into account by pre-commit
jaunty juniper
the way pre-commit works is a mystery to me, it looks like it runs its own installs of the tools it uses with the appropriate version from some cache, but the details are a mystery
tulip sleet
did you get your version up to date so your .pre-commit-config.yaml was the same? I don't know whether that helps
devout jolt
yeah it was a fresh checkout
tulip sleet
anyway, did it reformat now after pre-commit autoupdate or removing the cache?
devout jolt
okay after doing the following:
rm -rf ~/.cache
pip3 install pylint --upgrade
pip3 install pre-commit --upgrade
brew upgrade black
pre-commit autoupdate
pre-commit run --all-files
I get:
does that mean it worked? it doesn't match CI
oh I have unstaged commits now. wow I guess I never focussed on the fact it REWRITES my files
tulip sleet
let me force a re-run of the CI, in case it was the pypi issue
devout jolt
hold on
tulip sleet
yeah, usually you do git commit, and it prevents that if something is wrong
devout jolt
I must say, I very much dislike how black reformats some things but this isn't my house 🙂
tulip sleet
we don't like it sometimes too, but consistency is better. pylint is lot more of a nuisance
devout jolt
Agreed. Thank you all for the help. Sorry if I was cranky
tulip sleet
well, it was very much not jiving with your expectations
devout jolt
whelp, it passed CI so looks like my offerings were accepted
jaunty juniper
I have learned how to force black to leave my lists one item per line even when they are small (ending comma), so I feel like I have the power now
weary hill
I added a PR for an auto-deploy feature for Piku if anyone is interested in testing or reviewing that would be welcome.
manic glacierBOT
stoic rain
Working on WIO Terminal eRPC support, and I've found that busio.UART doesn't implement RTS/CTS yet. Anyone know if this is just a to-do thing or if there was a deeper reason? Based on the SAMD datasheets it seems like RTS/CTS flow control should be implementable as there's hardware support for it in the onboard USART.
manic glacierBOT
stuck elbow
Working on WIO Terminal eRPC support, and I've found that `busio.UART` doesn't i...
I think there was simply no need for it, so nobody worked on it.
stoic rain
I think there was simply no need for it, so nobody worked on it.
So I guess I've found a need (the WIO Terminal's SAMD51 uses a UART to talk to the RTL8720D co-processor and includes hardware flow control to help ensure that eRPC messaging works as needed) and may need to work on it 😉
lone sandalBOT
fleet hollow
Been working on adding pan to AudioMixer/MixerVoice. So far it is functional though the algorithm is very basic and may have room for improvement. I've only tested with stereo 16-bit audio on an RP2040 Pico so far. I may just put a conditional in there to only allow this functionality if the self->channel_count == 2 since I assume all channels are in a continuous alternating stream of 16-bit values packed into the 32-bit buffer which I haven't accounted for. Any thoughts on this? I'll be working on some deeper audio processing additions to CircuitPython for personal use, but figured this could be generally more useful in case there's a potential pull request here. https://github.com/dcooperdalrymple/circuitpython/commit/01c8bd673a0675b155645350d7aa131347cba40c
tulip sleet
Been working on adding pan to AudioMixer/MixerVoice. So far it is functional tho...
There are a bunch of open issues about audiomixer: https://github.com/adafruit/circuitpython/issues?q=is%3Aissue+is%3Aopen+mixer. I was thinking particularly about https://github.com/adafruit/circuitpython/issues/3210 w/r/t what you are doing, but there may be others of interest to you.
https://github.com/adafruit/circuitpython/issues/5136 has some long-term thoughts about audio. Audio support is in general very complicated and timing-error prone, and we think eventually it might be reworked and cleaned up, both internally and in the API.
tulip sleet
@onyx hinge gm - trying to think about #6246. The fix does work, in a slightly accidental way. The alternative is to disable the pullup in the AnalogIn constructor, and document that you should wait at least one second before reading board.BATTERY
onyx hinge
I think the pullup should almost certainly be disabled in AnalogIn
measuring across a resistive divider is very common, and the pull-up changes the ratio -- markedly, in this case
tulip sleet
the time constant on this particular divider is so long, do you think we should keep the never reset thing also. It works because the pin is not reset with a pull-up elsewhere
onyx hinge
someone with a ppk2 could determine the power impact of disabling just that one pull-up
surely it's pretty small
fathom sand
EMC Testmode for nRF52840
this is a duplication of my post in the Adafruit forum,
and in the issues of CircuitPython in Github,
maybe it is better located here...
I am almost ready with a project / product containing a ItsyBitsy nrf52840.
It is a braille keyboard for blind users to connect to IOS devices.
We initially want to build a small production run.
Regardless of the number of products brought to the market we need a CE certificate in Europe.
(Similar to FCC in the US.)
For CE RED and EMV testing, the laboratory tells us that we need to bring the nRF52840 into an EMC-Testmode
In the Nordic nRF52840 SDK documentation I found this:
https://infocenter.nordicsemi.com/index ... erial.html
Here are my questions:
Is anyone here aware of this mode?
Can this mode be "activated" from CircuitPython? (currently, I assume not!)
In an ideal world, I would like to see a library with a function settestmode(mode) that would activate the different modes.
After this call the nRF Chip would be probably unusable for CP until the next boot...
Can anybody help with creating such a library?
Or where would be the correct place to put such a function into the nRF52840 builds.
I think this would be "a pice of cake" for someone knows how to talk to the SoftDevice of this chip.
tulip sleet
EMC Testmode for nRF52840
this is a duplication of my post in the Adafruit foru...
could you paste the "infocenter" link again? It is truncated
fathom sand
tulip sleet
for this testing, does it need to be running CircuitPython at all? If it's just an EMF test, then could you write a small Arduino program or similar and load it with that?
onyx hinge
does it have to be part of the shipping device firmware? exactly what Dan is thinking..
tulip sleet
the on-board module itself is CE-certified, so does the whole system need to be re-certified?
fathom sand
No it does not!
Also a "firmware" which just does the test mode would be fine...
The lab tells me, that I need to bring a product which goes into "EMC Testmode Center Frequency unmodulated"
onyx hinge
nRF5_SDK_17.1.0 has an examples/dtm folder. It has hex files built for various of their devkits, but perhaps switching that around for the module on the itsybitsy is .. feasible.
tulip sleet
The lab tells me, that I need to bring a product which goes into "EMC Testmode C...
are they aware of the pre-certified module?
perhaps Nordic has a standard firmware load already built that can be loaded onto it with SWD. But then the rest of the device would not function
maybe that is OK
The Nordic sales engineering team may be able to help you with this
lots of hits here: https://devzone.nordicsemi.com/search?q=ce certification firmware#serp=2
fathom sand
Yes the module is certified, which takes away most of the problems.
But the integration onto a PCB and into a housing needs to be tested.
And yes I do have the certification documents of the module and the
Lab is aware of this.
tulip sleet
We do not use the nRF SDK, so linking in the DTM module to CircuitPython would be a major undertaking. A standalone SDK program makes more sense, I think.
(the SDK is not the SoftDevice, I think you know that)
fathom sand
Thank you for the link, and I understand what you say.
tulip sleet
since we overwrite the standard bootloader, you may need to restore that to run some other firmware, that is the main nuisance I see
fathom sand
since we overwrite the standard bootloader, you may need to restore that to run ...
Yes, I will need to compile with the SDK a "special" firmware and somehow get this firmware onto the ItsyBitsy.
For this, I will need to use a Segger J-Link programmer and connect to the pads on the back of the ItsyBitsy.
This alone will be "tricky" as the ItsyBitsy is soldered onto a keyboard PCB....
Because of this, I had hoped to be able to activate the Test-Mode from a CP or even Arduione configuration.
tulip sleet
perhaps you can extract the DTM lib into an Arduino program
tulip sleet
@onyx hinge I think I will make a 7.3.0-beta.1 after #6426 and #6226 are merged and the translation updates are merged. Anything you want to get in before that? Still open questions on several other PR's
it will take all afternoon for all the merges to complete anyway
onyx hinge
<@426200547024961539> I think I will make a 7.3.0-beta.1 after #6426 and #6226 a...
Not particularly, but I'll do a scan of PRs a bit later today. Is 6226 the pio wrap one?
tulip sleet
filesystem stub
tulip sleet
oh, I can do the PIO wrap one too, missed that
tulip sleet
absolutely fine
I want to get the 7.2.5 fixes in a beta, that's a high prio for making a new one
ok, will do
I started on release notes already
manic glacierBOT
fleet hollow
There are a bunch of open issues about audiomixer: <https://github.com/adafruit/...
That's good to hear. I'll keep what I'm working on as a private build or eventually move over to C++. I definitely see major opportunities to ramp up the capabilities of the CircuitPython audio engine, but that may be outside of the scope of the project.
tulip sleet
@onyx hinge I approved https://github.com/adafruit/Adafruit_CircuitPython_PIOASM/pull/35. Maybe this should finally be 1.0.0 ?
onyx hinge
Oh is it still pre-1.0? It should be 1.0 at least.
misty garnet
anyone consider adding %timeit to the REPL?
tulip sleet
anyone consider adding %timeit to the REPL?
that's an iPython thing, not in the regular REPL in CPython either. But there is the timeit module, which could be adapted to CircuitPython
misty garnet
timeit module looks pretty portable, just have to replicate the itertools.repeat function, which is pretty straightforward
proven garnet
For anyone that works with requests, here's a PR to allow for multiple cookies in the hrader. This seems to be in compliance with CPython but figured others might have thoughts or at least more knowledge. https://github.com/adafruit/Adafruit_CircuitPython_Requests/pull/108
My thoughts were that if the library is often frozen in, this would be a better strategy than implanting the cookies property or CookieJar class
warm flame
https://blog.adafruit.com/2022/04/06/circuitpython-7-3-0-beta-1-released/
What is usb host?
stuck elbow
the thing you connect usb devices to — usually your computer
jaunty juniper
supporting USB host allows connecting keyboards and other generic USB peripherals to a Circuitpython board
I believe AT Makers has a case of an accessibility device that uses a proprietary type of HID, that they want to convert into a generic compatible HID device, that's an interesting use
idle owl
@tulip sleet Can you take one more look at this PR? I'm blocked on it at this point for one of my tasks. If you're ok with it, please approve. If not, lmk what I can do. Thanks! https://github.com/adafruit/Adafruit_Learning_System_Guides/pull/2135
bitter marsh
I'm looking for someone to build me (then help me build) a "custom" CP image for RP2040 (Nano). It's really the same as, say, 7.2.4 but just with unused modules removed to help save heap space. Will pay of course. Thanks!
idle owl
I'm looking for someone to build me (then help me build) a "custom" CP image for...
Folks here will almost certainly help you build a custom CircuitPython image. However, we, as a community, don't typically accept pay to do this type of thing.
stuck elbow
I'm looking for someone to build me (then help me build) a "custom" CP image for...
do built-in modules take up heap space?
onyx hinge
looks like github actions is sad again today. such a variety of failures ... https://github.com/adafruit/circuitpython/actions/runs/2109289862
stuck elbow
I'm looking for someone to build me (then help me build) a "custom" CP image for...
have you already looked at https://learn.adafruit.com/building-circuitpython/build-circuitpython ?
onyx hinge
githubstatus is green, so they haven't noticed. https://www.githubstatus.com/
onyx hinge
do built-in modules take up heap space?
generally speaking, disabling a module frees up flash but doesn't affect heap much.
bitter marsh
have you already looked at https://learn.adafruit.com/building-circuitpython/bui...
I have and that would be the next logical step. Just trying to save myself some time. I find I sometimes spend more time setting up then actually doing.
bitter marsh
generally speaking, disabling a module frees up flash but doesn't affect heap mu...
That does make sense becuase it's not loaded.
stuck elbow
what is the problem you are trying to solve?
bitter marsh
what is the problem you are trying to solve?
Might be a fools errand, but looking at trying to save heap space. We're right on the edge. It may be that it does little to nothing because the modules are not loaded anyways.
stuck elbow
@bitter marsh I would look into precompiling as much as you can into .mpy files, that will save some memory
bitter marsh
<@944686897060520058> I would look into precompiling as much as you can into .mp...
Of our code, I presume you mean? Yea, all of the libs we use are .mpy files and yea, we'd need to look at our own. also I presume converting fonts to binary will save a decent amount too.
tulip sleet
<@329766224093249548> Can you take one more look at this PR? I'm blocked on it a...
Somehow I missed your response, sorry. Approved.
idle owl
Somehow I missed your response, sorry. Approved.
Ah ok. Thank you!
@tulip sleet I re-requested review on the .gitignore PR as well.
idle owl
@tulip sleet Hey... question. What should I call the asyncio template in a board guide? "asyncio"? something else? It goes under the "CircuitPython Essentials" header, so it doesn't need "CircuitPython" in it. But I'm not sure if you have a better idea for a page title, or if you think asyncio is fine.
tulip sleet
i was calling things "Cooperative multitasking", but that's a mouthful, and asyncio is the mechanism to do it and the shorthand, so go with asyncio
candid sun
hey @tulip sleet - i'm using the keypad library and trying to see if there's a simple way to detect a double-click. currently trying with time.monotonic() and states but wondering if i'm missing something with the event queue maybe that would be easier?
tulip sleet
hey <@!329766224093249548> - i'm using the keypad library and trying to see if t...
hi - there is a timestamp in each event: event.timestamp. There is a library for doing arithmetic on these, to handle wraparound. let me find it
candid sun
<https://github.com/adafruit/Adafruit_CircuitPython_Ticks>
ah cool, thank you!!
tulip sleet
courtesy jepler; he added the timestamps
idle owl
@tulip sleet Alright, the asyncio template is ready for your review. I'm not sure if you're able to edit templates or not. If you edit the guide page, there will be a link at the top to the template (like on mirrored pages). You would click that to go to the template editor, if you have access. Otherwise, I guess we can discuss your suggestions in realtime here, or we can open an issue on the Learn repo, if you'd rather. Up to you. https://learn.adafruit.com/adafruit-esp32-s2-feather/asyncio?preview_token=vJ_qAZ3jfU7F3eDEmDKYUA
Oh, you're not expected to copyedit (but you can if you want), it's more reviewing it for accuracy and so on.
A lot of it is from your guide, so it's probably ok, but you might have better ways to explain some things that I didn't get from your guide.
tulip sleet
I see "This page is templated, You may only edit content in the marked sections."
idle owl
You don't have this at the top of the page?
tulip sleet
No.
idle owl
OK, then you don't have template access. Which makes sense, only 4 people do I think.
tulip sleet
so I don't have the privs
Is the actual name of the page "asyncio", or that's the name of the template?
idle owl
Both. I can change the template name to be clearer if you think it would be helpful
tulip sleet
I think "Multitasking with asyncio" is probably less cryptic. "Cooperative multitasking with asyncio" is maybe too long?
idle owl
Ok, that sounds good. You choose the page title when you add the template, so the template name doesn't matter.
Title updated.
tulip sleet
Maybe put colons after the "one" and "two:
NeoPixel ring one data in (DIN) to microcontroller A1
NeoPixel ring one ground to microcontroller GND
NeoPixel ring one V+ to microcontroller 3V
NeoPixel ring two data in (DIN) to microcontroller A2
NeoPixel ring two ground to microcontroller GND
NeoPixel ring two V+ to microcontroller 3V
idle owl
Hmm. We don't usually do that. You think it might be clearer that way?
tulip sleet
"highlighted in [color] the wiring diagram" -> "on the writing diagram"
I thought it was "one data in" I parsed it wrong the first time I read it.
it's "NeoPixel ring one: data in ..."
idle owl
Fair enough. Ok. Fixing template now, then will fix guide page.
Fixed.
Fixed the text for the button too
idle owl
Ya I knew what you meant. Ya, used in 🙂
tulip sleet
"You should have at least the following file in / of the CIRCUITPY drive:", maybe say "top level". Novices may not know what "/" means
idle owl
Ooh, that's a Learn-wide thing, there. You're right though. I'll do that from now on. Fixing now.
Fixed
The "at least" wording is because of the sometimes superfluous adafruit_bus_device or pixelbuf.mpy that are included with some libs, even when not needed because the screenshot generator isn't smart enough to know when they are not needed.
If you download the project bundle, your lib will match the screenshot.
asyncio doesn't work on boards that would require adafruit_pixelbuf.mpy so it works here.
tulip sleet
"Following set up, you create a class called AnimationControls. This class enables you to include ways to control the animations for use with asyncio." => "... This class provides ways to control the animations ..." or "This class provides the ways the animation can be controlled ..."
tulip sleet
I think you could says something like "Notice that the controls object provides the animation direction (controls.reverse), the delay between steps of the animation (controls.delay), and the delay between complete animations (controls.wait). That would go below def rainbow_cycle or maybe def blink.
.delay vs .wait is a little confusing, one is intra-animation, the other is inter-animation (inside vs between)
idle owl
Hm. I can see that.
tulip sleet
e.g. this uses the word "wait":
You fill all the NeoPixel LEDs blue, wait for a specified amount of time, then turn all of the LEDs off, and wait for the same specified amount of time.
waiting in this case for .delay time
you could say "delay for a specified..."
tulip sleet
"The blink() block twice includes the following line of code which utilises the delay variable from the AnimationsControl class above." => "the blink() block includes the following line of code twice, which uses [or utilises] the .delay attribute of the AnimationsControl object." Similarly in next sentence, refer to the `.wait' attribute in the object. I like being precise about object vs class.
idle owl
I do too, but I'm fuzzy on it, so I don't get it right.
tulip sleet
and the technical name is attribute, rather than variable, because it's an attribute of the object. i.e. not a plain old variable
idle owl
Right that makes sense.
tulip sleet
"or the task function finished", can leave out the "or"
tulip sleet
"If you want to ensure the task executes forever, have a loop that never ends in the function" => "If you want to ensure the task executes forever, have a loop that never ends in the task function"
idle owl
need to update the second instance of variable.
idle owl
"If you want to ensure the task executes forever, have a loop that never ends in...
Also good call
There I think I got both of the places you meant.
tulip sleet
async def loop():
while True:
print("I'm looping!")
await asyncio.sleep(0)
I would call this task() instead of loop, or never_ending_task(), or something to emphasize it's an example of a task.
tulip sleet
"a loop in your task function"
tulip sleet
that's it; thanks for your patience
idle owl
Thank you so much! This was all great.
tulip sleet
I think it's a very clear example, and well laid out. yw
idle owl
Thank you! 😊
tulip sleet
"Both functions end with the following line of code which utilises the .wait attribute of the AnimationsControl class." class => object
tulip sleet
ok, made a second pass, looks great!
idle owl
I can hand it off to Anne and Limor with more confidence. Thanks so much for all of your input
tulip sleet
you're very welcome!
idle owl
@gilded cradle Hello! Are you around for a couple of questions? They're about the WebSerial ESPTool URL changes you made that came up last night in the meeting.
gilded cradle
<@478616919721771019> Hello! Are you around for a couple of questions? They're a...
Sure Kattni
idle owl
@gilded cradle OK. First question is the changes were made in the "Web Serial ESPTool" pages, as well as the "Factory Reset" template/pages, correct? Or are there more instances of it to look for?
@gilded cradle The second question is, are you going to set up https://adafruit.github.io/Adafruit_WebSerial_ESPTool/ with instructions to go to the other URL, or as a redirect, or something? Or will it simply remain as is.
gilded cradle
As far as I know I changed all the relevant ones. If you find any they should be updated.
idle owl
As far as I know I changed all the relevant ones. If you find any they should be...
Ok, I'll do a bit of digging to make sure, but sounds good.
gilded cradle
We can set it up to redirect
idle owl
Might be worth it. I have it bookmarked, and that's what I was using to get to it.
gilded cradle
I have to wait for a PR to get merged because Github doesn't like you having forks at 2 different locations.
idle owl
Ah fair enough. I've dealt with similar before.
gilded cradle
🙂
idle owl
Thanks!
onyx hinge
@tulip sleet maybe you can explain the philosophy & practicality of using #if FOO instead of #if defined(FOO) on this PR? I'm not sure I'd get the explanation right, because I'm comfortable with both ways of doing things and I'm not sure I'd get it right. (and maybe it should be a CIRCUITPY_ identifier too?) https://github.com/adafruit/circuitpython/pull/6261
tulip sleet
<@!329766224093249548> maybe you can explain the philosophy & practicality of us...
Sure.
idle owl
@proven garnet I figured out a way to generate a patch file for Adabot that only applies to Adafruit CircuitPython TestRepo, for the purpose of being able to test the process without actually trying to apply a patch across all the libs. This means I have a way to walk you through it to find out if you can do it without the pressure of actual patch contents. I think sometime next week will be best as I've got obligations the rest of this week to attend to.
Congrats to @proven garnet for submitting the 350th CircuitPython library! Between the Adafruit CircuitPython libraries and the Community Bundle, we've reached 350!
@gilded cradle Found one. https://learn.adafruit.com/admin/guides/3358/editor/23525 Can you update the screenshot? I already updated the URL, but realised the image didn't match anymore.
gilded cradle
Sure
idle owl
Otherwise, it looks like you caught everything!
gilded cradle
I updated like 3-4, so I think I got the process down.
idle owl
There's one more guide returning on the URL search, but I can't find a reference to it in the guide. Justin is checking.
I'm not editing, you can take over. I closed the tab.
gilded cradle
Ok, cool
idle owl
@gilded cradle Justin found it. https://learn.adafruit.com/upgrading-esp32-firmware/upgrade-rp2040-airlift-firmware#flash-airlift-using-web-serial-esptool-3086368-18 <-- also needs to be updated.
gilded cradle
@idle owl Ok, that one's a bit different and will need new screenshots.
idle owl
Yeah, that's what it looked like. Anyway. That is definitively the last instance of it. 🙂
gilded cradle
Thanks
gilded cradle
<@478616919721771019> Found one. https://learn.adafruit.com/admin/guides/3358/ed...
Ok, Kaluga one is done.
idle owl
Thanks!
gilded cradle
yw
proven garnet
<@135415127879974912> I figured out a way to generate a patch file for Adabot th...
That sounds great! I'd love to go over it then!
idle owl
That sounds great! I'd love to go over it then!
Excellent. We'll sort it out next week!
jaunty juniper
anybody has a Feather ESP32S2 TFT with latest CP ?
gloomy shuttle
anybody has a Feather ESP32S2 TFT with latest CP ?
I have one
dusk mauve
I have one as well, just finished doing a fresh build/flash, have some time this evening if you need someone to test something @jaunty juniper
timber mango
Question - I am new at circuit python so I am not very good at it and might be asking stupid questions.
The CAN library does not seem very good. It has all blocking calls instead of callbacks while stuff happens in the background
Can I create a library in C, use the microcontrollers built in CAN functions, and then create an interface for MP/CP?
Same for if I went a offbord CAN solution with SPI and DMA
The goal is 2 things: make an arduino interface and a circuit python interface (hopefully saving having to rewrite code), and to have higher performance CAN code that does not block
