Made it "".
#circuitpython-dev
idle owl
lone axle
hmm that was my first thought
idle owl
Right?
I thought I was so slick figuring it out.
Hold on, I can do it and give you the failure.
lone axle
okay cool, maybe it's because it tries to read the value later in one of the other prompts should be some way to get it going though.
idle owl
Oh come on.
It worked this time.
Well, almost. There's a space after Product Name. pid [0]: 4458 sensor_name []: Product Name directory_name [Product Name_pcb]: sensor_description []: product_name [Product Name ]: repo_name [Adafruit-Product-Name-PCB]: github_repo_url [git@github.com:adafruit/Adafruit-Product-Name-PCB.git]: product_copy [INSERT PRODUCT COPY HERE]: image_url [https://cdn-shop.adafruit.com/1200x900/4458-01.jpg]: eagle_file_directory []: designer [Limor Fried/Ladyada]:
How do I make the space exist only if there's a description?
I assume it's this "product_name": "{{cookiecutter.sensor_name}} {{cookiecutter.sensor_description}}", adding the space.
lone axle
yep I think you are right. Generally speaking conditionals work like:
{% if condition %}anything in here gets included if condition was true{% endif %}
so something like
"product_name": "{{cookiecutter.sensor_name}}{% if cookiecutter.sensor_description %} {% endif %}{{cookiecutter.sensor_description}}",
I think
oof the line breaks
idle owl
Trying it
I renamed some things, fixing all the usages now.
Got it! @lone axle Thank you so much. This will save me a little bit of time every time I do this. Cheers!
idle owl
lone axle
Nice you're welcome. happy to help out
idle owl
Doesn't seem like much, but oi, not having to deal with fixing that every time will help me a lot.
manic glacierBOT
idle owl
@lone axle Hah, apparently the same issue is present in the main cookiecutter. Choose from 1, 2 [1]: github_user [adafruit]: author_name: Kattni Rembor company [Adafruit Industries]: library_name: Simple Text Display library_description []: A helper library for displaying lines of text on microcontrollers with built in displays. library_keywords []: library_prefix [adafruit]: adafruit_pid []: requires_bus_device []: no requires_register []: no other_requirements []: pypi_release [yes]: no sphinx_docs [yes]: gives adafruit_simple text display.py
I don't even have a clue where to start fixing that. Wait..... is that what your PR fixes?
lone axle
I think it is one of the things.
there were several spots that ended up needing changed up a bit.
idle owl
Yeah I would assume so.
Hmm ok. Can you take a look at that PR? I pinged you on it anyway for other reasons. It looks mostly done?
lone axle
Yep, just noticed further updates on that. and realized I missed a hug report this week for patrick, will get it in next week.
idle owl
@lone axle I just tested your PR and ended up with an empty folder with .github in it. Am I missing something?
At least I'm pretty sure I checked out the PR. I used the gh command
lone axle
odd, I ended up with files inside of there when I run it
idle owl
Let me try again.
lone axle
lone axle
to run it you did gh pr checkout 140 and then cookiecutter path/to/cookiecutter-adafruit-circuitpython/ ?
idle owl
I ran cookecutter . for running it from inside the repo directory. Is that wrong?
and yeah that's the gh I ran.
lone axle
I tend to do it from one directory above so I end up running cookiecutter cookiecutter-adafruit-circuitpython/ and the new library gets created along side of the cookie cutter repo
idle owl
Ok let me try that.
Same empty directory.
@lone axle Wasn't there a second jinja directory causing issues?
Is it that it wasn't removed locally for me?
lone axle
Hmm yeah must be something else going on, I just verified I am able to use cookiecutter . as well and that does generate the files as expected for me.
Ah that is true when those PRs got crossed and it introduced a second one. I'm not sure how it decided which to use.
idle owl
This is my cookiecutter repo directory.
lone axle
still zephyr
Will do ๐
idle owl
@lone axle That made it work.
@lone axle I'm going to merge this. The README is fixed, I trust the rest already worked.
lone axle
sounds good I believe it should be all good now. if anything else pops up we can get it fixed up in a future PR still
idle owl
For sure.
manic glacierBOT
manic glacierBOT
jaunty juniper
huh releasing 6.3.0 didn't remove the rc0 ๐
back to the drawing board (not tonight though)
jaunty juniper
ah wait, that's because that PR si not part of the selection for the 6.3.0 update !
sigh of relief
still, need to manually remove the rc.0 now ๐
manic glacierBOT
twilit kiln
Hey @tulip sleet.I did coordinate with the Zephyr folks.They did share this link with me https://github.com/zephyrproject-rtos/zephyr/issues/22748, regarding the progress on the pinctrl api.They told me that it can be used for runtime pin assignment, because it makes it possible to construct the pinmux routing at runtime.Can you please have a look once.Like you told me previously that most of the work revolves around build systems, what can I do from now onwards for porting CircuitPython to Zephyr RTOS.
orchid basinBOT
manic glacierBOT
idle owl
@lone axle For cookiecutter, the .py file generated has two copyrights in it, Scott 2017, and the author, this year. Why is that? Scott has written a lot, but not every new library four years ago ๐
lone axle
<@!382939733107408897> For cookiecutter, the .py file generated has two copyrigh...
I don't think I've ever been told a specific reason. The conclusion I came to myself though was that Scotts copyright applies to the template, whereas the newer one with the new user applies to the new library file that has been created from the template.
still zephyr
@idle owl hello could you take a look at this PR, to see if you agree with the proposal. thanks ๐ Just wanted to validate with you https://github.com/circuitpython/CircuitPython_DisplayIO_SwitchRound/pull/4#issuecomment-853832947
idle owl
<@!330227457296957440> hello could you take a look at this PR, to see if you agr...
Done.
idle owl
Anyone have a PyPortal Titano they can test something for me? I know I have one but I'm struggling to find it.
orchid basinBOT
idle owl
Nevermind, found someone with a Titano.
idle owl
@gilded cradle You around for a quick discussion about a helper lib I'm working on?
I'm extracting SimpleTextDisplay out of the CLUE library into its own helper. https://github.com/adafruit/Adafruit_CircuitPython_CLUE/blob/f879c341765bcf0506013aa9037f91aaf6e673a7/adafruit_clue.py#L63 and https://github.com/adafruit/Adafruit_CircuitPython_CLUE/blob/f879c341765bcf0506013aa9037f91aaf6e673a7/adafruit_clue.py#L889.
frozen flume
How can I make this box appear on my board? https://circuitpython.org/board/atmegazero_esp32s2/
idle owl
@frozen flume It would have needed to be included in the 6.3.0 release.
There will be a 7.0.0-alpha.3 soon which will show up similarly. And is a direct download, so folks don't have to deal with S3.
frozen flume
I haven't seen it show up and it's been a while
I've seen a few releases and nothing
idle owl
The 6.x releases were a different branch. Dan's working on the 7.0.0 release right now.
There hasn't been a 7x yet which is why it hasn't shown up.
frozen flume
ah ok, that makes sense. Thanks for your quick reply ๐
idle owl
You're welcome! Dan's literally working on the release as we speak. I believe he's out for a few days, so if it doesn't get out today, it'll be next week. But that's when it'll show up.
@lone axle Display Text question, if you have a moment.
lone axle
<@!382939733107408897> Display Text question, if you have a moment.
yep, whats up?
idle owl
yep, whats up?
Label. Has max_glyphs. If you give it a number, and then enter more characters than that number, it should fail right? Or did something change in how that works.
Here I specify 45, but I can enter WAY more than that and it displays them all properly: https://github.com/adafruit/Adafruit_CircuitPython_CLUE/blob/f879c341765bcf0506013aa9037f91aaf6e673a7/adafruit_clue.py#L140
lone axle
Two things: BitmapLabel has mostly the same API and never did have the max_glyphs restriction, so generally I recommend folks to use that one instead of Label. But also that max_glyphs inside of Label was tied to the max_size inside of displayio.Group which did end up changing in the core and there is no longer a max_size restriction on Groups which means that Label will no longer fail if you go over the max_glpyphs value. We could probably remove max_glyphs all together from Label to avoid confusion.
idle owl
Ahhh.... ok
Right on.
On both counts.
I'll see about converting this to BitmapLabel
gilded cradle
<@!478616919721771019> You around for a quick discussion about a helper lib I'm ...
Sure @idle owl
lone axle
I'll throw it on my list to PR Label to remove that max_glyphs parameter.
idle owl
I'll throw it on my list to PR Label to remove that max_glyphs parameter.
Perfect thank you.
idle owl
Sure <@!330227457296957440>
I'm extracting SimpleTextDisplay out of the CLUE library into its own helper. https://github.com/adafruit/Adafruit_CircuitPython_CLUE/blob/f879c341765bcf0506013aa9037f91aaf6e673a7/adafruit_clue.py#L63 and https://github.com/adafruit/Adafruit_CircuitPython_CLUE/blob/f879c341765bcf0506013aa9037f91aaf6e673a7/adafruit_clue.py#L889.
lone axle
If everything is working as intended it should be as easy as changing the import and constructor over to BitmapLabel, everything else should work the same as with the regular Label.
idle owl
@gilded cradle Within the CLUE library, the imports for the text display are INSIDE the class because we didn't want them to tie up the display stuff unless you wanted it.
gilded cradle
I'm extracting SimpleTextDisplay out of the CLUE library into its own helper. ht...
Ok, I'll taking a look...
idle owl
But since I'm extracting it into its own library, I can just import at the top right? Because you have to have those things if you're using this new library.
So there's no point in doing the funky inside-the-class-imports.
idle owl
If everything is working as intended it should be as easy as changing the import...
Oh nice. OK. Thanks
gilded cradle
Yeah, that's what I did with the FakeRequests library
idle owl
Yeah, that's what I did with the FakeRequests library
Right on, ok. Simply wanted to make sure I wasn't missing something. Thanks!
gilded cradle
Right on, ok. Simply wanted to make sure I wasn't missing something. Thanks!
Nope, it should be pretty straightforward.
The hardest part is just getting all the new library framework all set up.
idle owl
Yeah, and tweaking this for other updated code.
idle owl
If everything is working as intended it should be as easy as changing the import...
I did from adafruit_display_text import bitmap_label as label and everything works. Is that sufficient?
lone axle
yep that works as well. if you rename it in the import like that then there is no need to update any of the rest of the code.
idle owl
Right on. I tried renaming it everywhere and it got sad. So I did the import trick instead.
Wanted to make sure it was actually doing what I thought it was.
idle owl
@lone axle board.DISPLAY is only available for boards with built in displays, right? And if you connected a display to a board, it would involve multiple pins. But you could create a display object with multiple pins and then feed it into something else as a single object, right?
Trying to figure out if this simple text display thing can be used with non-built-in displays.
lone axle
yep, I tend to do it with renaming like that so that I can easily switch code back and forth to test both types of labels. So that is definitely a correct way to do it.
That is correct board.DISPLAY won't exist on devices without a built-in screen. Instead the user would set up the Fourwire and end up creating a display object using the driver matching their hardware.
idle owl
But the FourWire thing ends with a single display object?
so if there's a display= kwarg, you could feed it that?
lone axle
yep, that is correct and that is probably how I would approach it as well.
idle owl
Ok
lone axle
https://github.com/adafruit/Adafruit_CircuitPython_DisplayIO_SSD1306/blob/master/examples/displayio_ssd1306_simpletest.py here is a good example that sets up the external display
idle owl
Ah thanks.
lone axle
line 34 creates the display object and once that object is created you call show() on it just like you would board.DISPLAY for a built in one.
idle owl
excellent.
tulip sleet
FYI, 7.0.0 alpha.3 is released. Blog post scheduled for 2pm EDT. circuitpython.org will update as soon as its job runs. Uploads are all there (except for one stuck job; the duplicate run that always happens will cover that).
idle owl
FYI, 7.0.0 alpha.3 is released. Blog post scheduled for 2pm EDT. circuitpython.o...
Thanks so much, Dan!
slender iron
thanks for the release @tulip sleet !
idle wharf
@slender iron may I join the Deep Dive thing (role\group). Reminders sound great !
slender iron
<@!252717193496756235> may I join the Deep Dive thing (role\group). Reminders s...
done! thank you!
idle wharf
Thank me... you do all the diving! Thank YOU !
ionic elk
@lime trellis @tulip sleet off this week but we can definitely start an issue for SAMD low power. The new internal structure should make it easier for community contributors to help, too. I imagine SAMD features will probably be closest to STM32, as opposed to more unusual implementations like NRF or the RP2040
slender iron
anyone with an m0 bluefruit want to help test: https://github.com/adafruit/Adafruit_CircuitPython_BluefruitSPI/pull/20
slender iron
<@!492049388152029185> <@!329766224093249548> off this week but we can definitel...
Issue is here: https://github.com/adafruit/circuitpython/issues/4837
blissful pollen
During "in the weeds" a couple weeks ago discussing the FunHouse display it was mentioned it would probably need a saleae to figure the timing out. Been thinking of getting one but wondering is the basic Logic 8 probably enough or worth the extra to get the Pro? Not sure if anyone working on CP has an opinion on it?
orchid basinBOT
idle owl
During "in the weeds" a couple weeks ago discussing the FunHouse display it was ...
Scott's worked with both, afaik. I have the 8 and it's been enough. I think the Pro goes to higher speeds, so if you're in a situation to need that, get the Pro. Otherwise the 8 is excellent.
blissful pollen
Thanks, yeah I saw the speed was the main difference and not sure if I'd ever need it, but I'd hate to buy it and a month later wish i'd go the better one
still zephyr
@ember iris ๐
ember iris
How's it going haha
still zephyr
ok so the any reference
sphinx will tell you the number of the problem as normal python code does, however sphinx only counts from the start of the """
so adafruit_sgp40.SGP40:29: will be in line 29 of the docstring
probably you mentin somenthing enclosed with ....
with backsticks and sphinx could not find the reference to that, sphinx will look in the function, the class documentation, and lastly the interphinx mapping
to look for the reference enclosed in the backsticks.
ember iris
Ok, based on that I'd guess it's upset about,
(...) you can access the class function `sgp.measure_raw` for a humidity compensated raw reading
still zephyr
so easy fix, would be to enclosed in doble backsticks
yes
if this is a function
you need to do it like this :meth: ` measure_raw
and end with another backstick
:meth:`measure_raw`
as mentioned before sphinx will look for in the class for that method so the sgp most of the time is not needed
ember iris
Ok, I'll do that and drop the leading sgp.
still zephyr
yeah try, if everything fails you could use double backsticks, it is accepted in sphinx, and in our design guide ๐
the difference, is for the user when you use the :meth: method the user could clik and go to the function definition directly, with double backsticks they could not
ember iris
Huh, I didn't know that. That's really nifty
still zephyr
yes ๐
ember iris
Ugh I need to set aside a weekend to just take a crash course in sphinx
but learning experiences are worth it
๐
still zephyr
Happy to help.. sphinx has a steep curve, and sometimes does not behave as expected...
manic glacierBOT
ember iris
Happy to help.. sphinx has a steep curve, and sometimes does not behave as expec...
Thank you for this help, this part of the pull request--the documentation--was honestly most of what felt daunting
and woo on my end the checks have passed
still zephyr
Thank you for this help, this part of the pull request--the documentation--was h...
Nice, Good ๐
ember iris
Nice, Good ๐
Woo and they passed on the pull request as well!
Ok
So, on my end I need to remove the register library from the requirements, and then what else needs to change--what am I looking for to see what I've missed and need to change before the pull request can be completed
still zephyr
Woo and they passed on the pull request as well!
Ok
So, on my end I need to remo...
Not sure I would do a review probably tomorrow, and I will put my comments
ember iris
Fantastic! And thank you for all of your help through this. I'll get the register reference removed then push that commit this evening so it should all be ready, plus any loose ends I notice
still zephyr
Fantastic! And thank you for all of your help through this. I'll get the registe...
yes no problem, is the register library also in the sphinx_intermapping, maybe you could verify the conf.py in the docs directory to check
manic glacierBOT
ember iris
yes no problem, is the register library also in the sphinx_intermapping, maybe y...
will do. There's not a call to it in the sgp40.py file so I'm decently confident that it's not used, but I'll walk through every file I can to make sure
still zephyr
will do. There's not a call to it in the sgp40.py file so I'm decently confident...
Thanks, I will take a look also
slender iron
During "in the weeds" a couple weeks ago discussing the FunHouse display it was ...
I'd recommend the Pro for the higher sampling rate. It means you'll be able to monitor faster signals. Have you thought about doing GitHub sponsors? I'd give a little to help offset the cost.
manic glacierBOT
blissful pollen
I'd recommend the Pro for the higher sampling rate. It means you'll be able to m...
Thanks I'll take a look at it, as it wasn't that much more. I hadn't thought about sponsors, I appreciate it. I'm not sure if I will or not right now. Long story short I finally did my tax refund from two years ago and got some back
slender iron
I think most people get an amount that can help ease getting the nicer thing rather than quitting their day job. ๐
blissful pollen
Good point. I'll have to look into how I can set it up.
manic glacierBOT
Looking into it a bit more, do we have evidence this is a net improvement in the build time? I compared the run for this build with a main-branch build.
The run for this PR says a slightly lower (better) number for "total duration" but if you add the times for the "pre...
manic glacierBOT
manic glacierBOT
hollow gazelle
What hardware did you use to support your RS485 needs? I don't seem to find any hardware on @river quest website.
fallen anvil
@hollow gazelle wow, that's a bit of a necropost! That particular design was custom and had a driver chip on it, but if you've already got rs232 you can use something like this to do the job; https://www.ebay.co.uk/itm/272345452345
There are plenty of other similar ones around, I don't know that specific board or seller so don't treat it as a recommendation.
manic glacierBOT
stone iris
Using circuit python 6.20 (have tried 6.30) getting an issue with microcontroller.nvm on a Raspberry Pi Pico. Can read bytes with no problem but when writing a bytearray either in code.py or in the repl then the system hangs and have to disconnect the Pico. When reconnected a read is showing the bytes have been correctly written. Has anyone an idea of a workround or solution. Thanks
blissful pollen
@lone axle I noticed on your stream you used VSCode, someone on twitter was asking about setting up environments other then Mu, did you have any links/tutorials to set up VSCode for CP? (or anyone else who reads this too)
analog bridge
Using circuit python 6.20 (have tried 6.30) getting an issue with microcontrolle...
@stone iris Can you share your code.py... I did a quick test with following running on pi pico with CP 6.3 and it works fine for me.
@slender iron I just discovered the deep divers role... can you add me to it.
jaunty juniper
<@!511885017887014933> Can you share your `code.py`... I did a quick test with f...
I'm doing tests and it takes writing more than once sometimes but I can repro except... sometimes
analog bridge
import microcontroller
microcontroller.nvm[0] = 1
print(microcontroller.nvm[0])
that can be enough:
>>> import microcontroller
>>> microcontroller.nvm[0] = 0
sometimes it takes doing it twice
slender iron
What hardware did you use to support your RS485 needs? I don't seem to find any...
I have a rs485 feather design I use to talk to my tv and stereo
slender iron
<@!252717193496756235> I just discovered the deep divers role... can you add me ...
done!
analog bridge
sometimes it takes doing it twice
hmm... I haven't noticed that behavior...
I am doing another test now and trying to fill the complete array, maybe that will make a difference
analog bridge
done!
thanks for adding me scott and also a well deserved hug report for the deep dives 
๐
analog bridge
@jaunty juniper I did a full re-write... still unable to replicate that behavior.
lone axle
<@!382939733107408897> I noticed on your stream you used VSCode, someone on twit...
I use PyCharm mostly. No experience wiyh vscode sorry
blissful pollen
I use PyCharm mostly. No experience wiyh vscode sorry
No problem. Thanks I'll pass that as an environment choice along too
idle owl
<@!382939733107408897> I noticed on your stream you used VSCode, someone on twit...
There's a plugin or extension or whatever VS Code calls them for CircuitPython that probably has instructions. I haven't dealt with it, but apparently it's nice.
I also use PyCharm, but it doesn't feel like it's designed for a CP workflow, you kind of have to massage it, or make your own workflow that works in PyCharm.
Atom has a plugin as well. That I haven't done much with.
(I enjoy PyCharm.)
blissful pollen
There's a plugin or extension or whatever VS Code calls them for CircuitPython t...
Oh cool thanks. I'll pass that along as well! I think I installed the CP extension for VSCode but never used it
analog bridge
Using circuit python 6.20 (have tried 6.30) getting an issue with microcontrolle...
I think I get the issue you are facing... are you trying to write multiple bytes individually?
If yes, then I would recommend the following method...
microcontroller.nvm[0:3] = b'\x00\x01\x02'
writing multiple bytes individually is very inefficient as it involves a complete flash erase each time a new byte is written...
the above method writes multiple bytes at once so only one flash erase is required
idle owl
@candid sun Hey! I'd like to help you add a page template to your guide for the install.
still zephyr
@ember iris ๐
ember iris
I was just about to grep through an see if there's a trailing reference to busio then ping you
still zephyr
@ember iris do you want to trade places?
๐
please?
could you review this PR? https://github.com/adafruit/Adafruit_CircuitPython_SHTC3/pull/11
ember iris
haha I can certainly try, but I'm still getting a feel for this so a second set of eyes past me would help too
still zephyr
this one is documentation only, I would like fresh eyes to see if I missed something ๐
ember iris
Will do, let me take care of this really quick then I'll review it. Fair warning I'm a dreadfully slow reader so it might take a bit
still zephyr
Thank you very much, really appreciated
idle owl
@slender iron Are the CP reports supposed to be running as you? Running GitHub checks as tannewt
slender iron
probably not ๐
idle owl
Didn't think so.
still zephyr
@idle owl hello
idle owl
Hello
idle owl
@still zephyr We can continue any discussion here.
still zephyr
So, I think the problem is in here
still zephyr
htmlhelp_basename = "Adafruit_CircuitPython_Simple_text_display_Librarydoc"
candid sun
<@!347203168121126912> Hey! I'd like to help you add a page template to your gui...
hi! thanks, i'm free the rest of the afternoon ๐
still zephyr
it needs to be comparing with all the rest of the librarieshtmlhelp_basename = "Adafruit_CircuitPython_Simple_text_displayLibrarydoc"
The other libraries does not use the underscore
idle owl
hi! thanks, i'm free the rest of the afternoon ๐
Great! Do yo you went to vid chat? Might be quickest.
idle owl
The other libraries does not use the underscore
Ah ok, well give it a try. We can merge it and see what happens! It's a new lib so we have room to mess with it.
candid sun
Great! Do yo you went to vid chat? Might be quickest.
yes sounds good!
still zephyr
Ah ok, well give it a try. We can merge it and see what happens! It's a new lib ...
Will do Thanks (y)
still zephyr
Thank you @ember iris 
ember iris
Thank you <@!459119350851567626> <:neokey:844950639250898955>
Woo! Thanks for asking!
raven canopy
<@!252717193496756235> Are the CP reports supposed to be running as you? `Runnin...
(& @slender iron) when run on Actions, the best i could determine in the past was that the "user" is determined by the person that made the last merge. my guess is that the git user is derived via the github.actor env var in the Actions runner instance.
slender iron
I think it's based on what API key the script uses
raven canopy
that could be true as well. is adabot's key not separate in the repo, though?
still zephyr
@idle owl PR is ready for the simple_text.
hollow gazelle
I have a rs485 feather design I use to talk to my tv and stereo
is that rs485 feather an adafruit product? - here's my $$$ ๐
idle owl
<@!330227457296957440> PR is ready for the simple_text.
Merged.
idle owl
Hmm.
still zephyr
dahs_display built without any problems, so it has to be something after.
idle owl
The docs are there. It's the subproject that's the issue it seems like.
still zephyr
yes the docs are perfect, but somehow readthedocs could not find the link
idle owl
It's my fault.
Hold on.
I named the slug wrong.
It's fixed.
I'm sorry for sending you down a rabbit hole!
still zephyr
๐
still zephyr
just for my info what is the slug?
idle owl
OK, so first you import a project into RTD. Then you have to add it as a subproject to the CircuitPython RTD project.
When you add it, you have to give it the doc name, which is what determines the URL.
I put in simple-display-text instead of simple-text-display.
So the URL was swapped.
idle owl
That's why it was building and passing, but the URL was 404. I should have checked that first.
slender iron
is that rs485 feather an adafruit product? - here's my $$$ ๐
its rs232 not 485 (my bad): https://oshpark.com/shared_projects/cPoaWCOm
idle owl
The circuitpython URL specifically.
still zephyr
I understand, well that is good, now we know
manic glacierBOT
hollow gazelle
<@649453464988155906> wow, that's a bit of a necropost! That particular design w...
Thanks, I ran across that - I was hoping to find some RS485 hardware that directly runs circuit python
today I've got my PC talking to one of these - maybe time to try blinka?
https://ftdichip.com/products/usb-rs485-pcb-pcb-only/
jaunty juniper
oh hey now that we have an alpha release, there should probably be a link to the 7.x bundle on https://circuitpython.org/libraries
still zephyr
@jaunty juniper the comment in this PR https://github.com/adafruit/Adafruit_CircuitPython_ESP32SPI/pull/115#issuecomment-853721812. Are you in favor of the changes, is that a correct assumption?
jaunty juniper
<@!199811382793863168> the comment in this PR https://github.com/adafruit/Adafru...
yeah, I think the changes make sense, upon tests it works, though the library itself is not very solid but that's a different problem
jaunty juniper
it really only uses parse_headers from requests, (which is not even in the requirements.txt apparently)
still zephyr
Yes I remember that much while testing Anecdata PR
hollow gazelle
its rs232 not 485 (my bad): https://oshpark.com/shared_projects/cPoaWCOm
no worry, in my case, the RS485 is needed for industrial communication, but in searching the discord and forums, it seems it is also useful for robotics.
( so far, it appears that even though we have CAN feathers, we don't have a RS485 feather or hat. - need to get something from sparkfun and get a soldering iron )
jaunty juniper
I think I get the issue you are facing... are you trying to write multiple bytes...
I am able to reproduce... something... every time now:
- nuked and flashed a pico with CPY 6.3.0
- open code.py, modify and save
- connect to REPL and do
import microcontrollerandmicrocontroller.nvm[0] = 5 - hang
Can you repro that ?
maybe even ctrl-C ctrl-D, just any soft reboot ? not sure
stone iris
@analog bridge @jaunty juniper positive it was all working a couple of days ago but microcontroller.nvm is changing whether writing a single byte or a byte array as per the code example given
Hanging not changing. The bytes are being written though.
tulip sleet
<@747673663419187281> <@199811382793863168> positive it was all working a couple...
Could you open an issue with your example? Thanks.
analog bridge
I am able to reproduce... something... every time now:
- nuked and flashed a pic...
@jaunty juniper I replicated the steps... still unable to reproduce the issue.
jaunty juniper
uggggh computers
twilit kiln
Hey @tulip sleet.I did coordinate with the Zephyr folks.They did share this link with me https://github.com/zephyrproject-rtos/zephyr/issues/22748, regarding the progress on the pinctrl api.They told me that it can be used for runtime pin assignment, because it makes it possible to construct the pinmux routing at runtime.Can you please have a look once.Like you told me previously that most of the work revolves around build systems, what can I do from now onwards for porting CircuitPython to Zephyr RTOS.
tulip sleet
Hey <@!329766224093249548>.I did coordinate with the Zephyr folks.They did share...
@slender iron had proposed having a meeting with you and your co-mentors to discuss the motivations and scope of this project. I donโt think that happened yet, is that right? Early next week would be necessary for this due to some scheduling constraints.
manic glacierBOT
twilit kiln
<@252717193496756235> had proposed having a meeting with you and your co-mentors...
Yes I will have a word with them and coordinate accordingly with you regarding the meeting.Thank you @tulip sleet .
manic glacierBOT
manic glacierBOT
twilit kiln
@slender iron can we please have a meeting tomorrow regarding the CircuitPython port for Zephyr RTOS.Thank you !!
slender iron
<@!252717193496756235> can we please have a meeting tomorrow regarding the Circu...
You need to email me (scott@adafruit.com) and your mentors to find a time that works for us all. I'm on vacation starting Thursday as well. Your mentors need to be there so they can understand the steps as well. They need to be the first place you seek help. We don't have the time to give the detailed help you've asked us for.
twilit kiln
You need to email me (scott@adafruit.com) and your mentors to find a time that w...
Sure @slender iron,I absolutely understand.Thank you very much for giving us your time.Lets finalise the time and day on mail then.
manic glacierBOT
manic glacierBOT
turbid radish
?serverinfo
digital shoreBOT
adafruit
orchid basinBOT
manic glacierBOT
idle owl
If anyone has a STEMMA QT Rotary Encoder breakout, I'd appreciate some assistance with testing something. Something isn't working and I'm not sure why.
gloomy shuttle
If anyone has a STEMMA QT Rotary Encoder breakout, I'd appreciate some assistanc...
I have one and can test something
idle owl
I have one and can test something
Or maybe I'm missing something. But the color changing works. The brightness changing when the switch is pressed does not. It should. Brightness works in seesaw, I tested it separately.
Also it's seeing the switch properly.
I have prints in weird places for testing, sorry, heh.
thorny jay
Dear CircuitPythonista, have not done any project, PR, successful test or anything. Except for for a group hug in hug report... I don't have much to say. So I will skip the meeting, maybe without notes.
idle owl
Dear CircuitPythonista, have not done any project, PR, successful test or anythi...
Ok. Thanks for letting us know. You're always welcome even when you don't have updates.
gloomy shuttle
Also it's seeing the switch properly.
To update brightness you want the button to be held down while rotating the encoder?
idle owl
Yes.
It's kind of annoying, but that's what the Arduino and Rotary Trinkey examples do.
so ๐คทโโ๏ธ
idle owl
To update brightness you want the button to be held down while rotating the enco...
Did I miss something obvious?
seesaw confused me a bit to begin with.
gloomy shuttle
Did I miss something obvious?
Your code looks fine, but I believe I am reproducing the same issue your seeing.
the brightness isn't updating
idle owl
Hmph.
blissful pollen
Hmph.
Long shot but does the neopixel brightness parameter work? I vaguely recall something about how it was supposed to be set at construction and not updated. I may be totally off here
idle owl
Long shot but does the neopixel brightness parameter work? I vaguely recall some...
I tested it alone, and it worked. I'm able to set brightness, and it changes.
That was my thought too.
Maybe it can't be changed on the fly through seesaw?
That would be odd.
blissful pollen
Yeah nothing is jumping out at me that is wrong. More so since the color is working
idle owl
If it can be set at all, it should be able to be set.
tidal kiln
@idle owl what's your code snippet that worked?
idle owl
And no, I don't need to call show() @tidal kiln ๐
idle owl
<@!330227457296957440> what's your code snippet that worked?
You mean what set the brightness properly?
tidal kiln
yep
tidal kiln
or just try via repl
tidal kiln
are you doing additional pixel setting in the loop?
tidal kiln
yah
tidal kiln
it looks like brightness is only factored in when setting something:
https://github.com/adafruit/Adafruit_CircuitPython_seesaw/blob/5cda48bcd2702737ec24ee5d4d2a921c050a4797/adafruit_seesaw/neopixel.py#L102
idle owl
Hmmmmmmm
idle owl
Right
tidal kiln
so maybe nothing is updating
idle owl
It's only brightness when pressed
ok
So it can't be updated on the fly through seesaw without color also changing or whatever.
Thanks for finding that. I would not have figured that out.
tidal kiln
i think so...looking at library code...
tidal kiln
while True:
position = encoder.position
if position != last_position:
# change the brightness
if position > last_position: # increase
pixel.brightness = min(1.0, pixel.brightness + 0.1)
else:
pixel.brightness = max(0, pixel.brightness - 0.1)
pixel.fill(0xADAF00)
last_position = position
try that? and just call fill() before loop with some color
idle owl
will try in a bit. meeting time.
gloomy shuttle
if you call fill in the loop it works
idle owl
Doesn't that mean it switches to that color while changing it?
hmm.
Is that worth it?
I guess maybe folks will want to know how to do it. So I guess so.
tidal kiln
yah. that's just a hack to test brightness behavior.
seems like sort of maybe a minor bug in the library? the fact that the brightness setter doesn't actually take effect immediately.
idle owl
@tidal kiln Would you want the hacky version or not have the brightness at all?
tidal kiln
looks like you're just trying to come up with a guide example?
idle owl
Yeah it's a quick example for the guide.
Because using the NeoPixel isn't intuitive.
The Arduino version already uses the neopixel.
But this will go in the seesaw library examples.
tidal kiln
looks like arduino version doesn't do brightness:
https://learn.adafruit.com/adafruit-i2c-qt-rotary-encoder/arduino
just prints "button pressed"
could do that?
gloomy shuttle
it does say this in the seesaw library "# TODO: brightness not yet implemented."
idle owl
I swear I just looked at the Arduino version and it controls the NeoPixel.
tidal kiln
yah, and skip brightness - or just set it once up top
idle owl
We are not looking at the same code.
https://github.com/adafruit/Adafruit_Seesaw/blob/master/examples/encoder/encoder_basic/encoder_basic.ino this does NeoPixel??
Or is wheel included for no reason?
Oh wait I get it
tidal kiln
looks same?
idle owl
doesn't do brightness.
tidal kiln
yah, right, just sets it once
idle owl
Ok, skip brightness.
tidal kiln
but then no change after that
void loop() {
if (! ss.digitalRead(SS_SWITCH)) {
Serial.println("Button pressed!");
}
int32_t new_position = ss.getEncoderPosition();
// did we move arounde?
if (encoder_position != new_position) {
Serial.println(new_position); // display new position
// change the neopixel color
sspixel.setPixelColor(0, Wheel(new_position & 0xFF));
sspixel.show();
encoder_position = new_position; // and save for next round
}
// don't overwhelm serial port
delay(10);
}
just color
idle owl
Thank you @gloomy shuttle, @blissful pollen and @tidal kiln . Really appreciate the help.
tidal kiln
that TODO note - i guess ideally the brightness math would be taken care of in the seesaw firmware. just hasn't been implemented yet. so, for now, it's in the library:
https://github.com/adafruit/Adafruit_CircuitPython_seesaw/blob/5cda48bcd2702737ec24ee5d4d2a921c050a4797/adafruit_seesaw/neopixel.py#L125-L130
idle owl
@tidal kiln
tidal kiln
yah
idle owl
Left brightness out. No need for it really. It works identical to neopixel so easy to do.
tidal kiln
is it crazy bright? the arduino example sets it to equivalently 0.2 in CP
idle owl
Well it's reverse mount, so unless you have it on one particular angle, it's sort of shielded.
eh I'll add it.
Put it to 0.5
Thanks for all the feedback.
tidal kiln
or...might be 20/255?
meh. yah. just some value. having the line there with any value will let people change it as they wish.
idle owl
Exactly.
idle owl
brb
slender iron
me too, water time
turbid radish
Lurking
sterile bronze
lurking
modern wing
Likewise, lurking happily ๐
onyx hinge
hmmm who has at least one thumb and didn't write their notes? ๐ this guy
idle owl
hmmm who has at least one thumb and didn't write their notes? ๐ this guy
Yeah I totally forgot too. Scrambling now.
modern wing
Yesterday? ๐
blissful pollen
Vacation is never soon enough
modern wing
Didn't really hear it, seems like the noise gate is good
onyx hinge
volume's a bit low @lapis hemlock
lapis hemlock
@onyx hinge @modern wing Thanks!
modern wing
lapis hemlock
Didn't really hear it, seems like the noise gate is good
What is a noise gate?
Is there a noise gatekeeper?
onyx hinge
Next meeting date change: July 4th's meeting is postponed to July 5 at the same time of day.
modern wing
@lapis hemlock An audio setting essentially truncates any loud/excessive noise in one's audio -- it can be a hardware or software tool. And as for the Gatekeeper, there's only Zuul.
onyx hinge
you can sign up for those hot hot rp2040 chips ๐ https://www.adafruit.com/product/5042
manic glacierBOT
trim elm
Next few days probably, only 90 libraries left to go!
blissful pollen
Is there a process to "take over" an older PR? Just comment and ask if the original author is about?
raven canopy
late...and lurking
onyx hinge
hey @raven canopy good to see you
raven canopy
hey @onyx hinge (and all!)
blissful pollen
Okay thanks, I'll take a look. The GH stuff I can probably figure out eventually
onyx hinge
you can check that the docs have been kept up to date with the code ๐ it's easy to forget/omit as a pull request author
lapis hemlock
onyx hinge
@gilded cradle random question, is "U2IF" counted as a single board in Blinka, or would each different U2IF board be a different Blinka board?
gilded cradle
@onyx hinge I'm not sure. We probably should count it for each board.
modern wing
So you're fixing everything, including the kitchen sink? ๐
onyx hinge
no, only the kitchen sink. everything else is in the same state as it was before
aha I think this is the crank -- an oldie, in effect
lapis hemlock
@idle owl you know that you can cut through the waves with a seesaw, don't you?
modern wing
Thanks everyone ๐
onyx hinge
Thanks everyone!
gilded cradle
Thanks
lapis hemlock
So long, everyone!๐
raven canopy
thanks all ๐
solar whale
๐
turbid radish
@fathom raptor, I'll finalize the newsletter once I get the PR
slender iron
๐
idle owl
All of the CircuitPython libraries have been moved to main!
modern wing
Indeed, enjoy your time off -- especially after this past year of insanity.
slender iron
I just switched the notes repo to main too
onyx hinge
nice, that should cause a minimum of trouble for me, I usually just use the web interface to updatei t
slender iron
yup, me too
solar whale
Finally, a good excuse to delete all my forks and get fresh clones....
gloomy shuttle
@idle owl I just submitted a pull request in an attempt to fix the brightness bug in Adafruit_CircuitPython_seesaw.neopixel
idle owl
<@!330227457296957440> I just submitted a pull request in an attempt to fix the ...
Oh nice!
I'm currently working on adding the example. I can pull down your PR and test it. I think RP2040 can handle that as a .py file.
idle owl
<@!330227457296957440> I just submitted a pull request in an attempt to fix the ...
Merged and released. Thanks so much! Saved me filing an issue. ๐
@gloomy shuttle Disconnected you from audio chat so nobody shows up and startles you.
slender iron
Here is the notes document for Mondayโs 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 youโll be attending the meeting - itโs super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and weโll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/10z3OQ1_Pc-cmG-CodmMsbFaebYaV81JSUGb8Tl-ql5o/edit?usp=sharing
slender iron
Here is the notes document for Mondayโs CircuitPython Weekly meeting. It is at t...
idle owl
OK, folks, the next time circuitpython.org builds, there will be a 7.x bundle available for download. We can stop sending folks to the bundle repo to get the 7.x bundle. FYI. Thanks for all your patience with us as we got to a 7.x release!
idle owl
@ember iris Be sure to let us know when you're ready to join the review team! You're always welcome to ask questions about how things are done, I know you were concerned about that with the documentation review you did.
@lone axle Good evening! Or afternoon or whatever it is where you are. Are you around for some cookiecutter talk?
lone axle
<@!382939733107408897> Good evening! Or afternoon or whatever it is where you ar...
on a call atm, but am available afterward. likely 20-30 minutes or so but not 100% certain
idle owl
on a call atm, but am available afterward. likely 20-30 minutes or so but not 10...
Ok, no worries. I might be around or not. Please ping me regardless and I'll let you know my status. Thanks!
ember iris
<@!459119350851567626> Be sure to let us know when you're ready to join the revi...
I'd be up for joining soon! I am busy today through wednesday but I've really enjoyed all of this!
idle owl
I'd be up for joining soon! I am busy today through wednesday but I've really en...
No worries at all! And no rush whatsoever. We'll be here when you're ready. And we understand life happens, so you don't need to be concerned about taking time for yourself.
ember iris
Awesome and thank you! Yeah just some summer chores the next few days which demand attention before the hot weather gets to the midwest
idle owl
It's already here. So I feel you there.
lone axle
Ok, no worries. I might be around or not. Please ping me regardless and I'll let...
Just wrapped up with that call. Am available now if you are around still.
idle owl
Just wrapped up with that call. Am available now if you are around still.
Let me go start laundry I should have started earlier, then I'd like to chat.
idle owl
@lone axle Ok. I'm creating a Community Code of Conduct to have in cookiecutter that goes into a lib if "Community" is chosen. I copied the current CoC, and then renamed both of them to basically this: {% if cookiecutter.target_bundle == 'Community' %}CODE_OF_CONDUCT.md{% endif %}
Is that even close to right?
Because I assume it has to have two files with names like that to pick one based on the bundle type.
I'm guessing entirely based on the other files though.
lone axle
<@!382939733107408897> Ok. I'm creating a Community Code of Conduct to have in c...
Yep, I do think that is right. using the conditional in the filename like that will make it so that the file does exist if the condition was true, and doesn't exist if the condition was false.
idle owl
Yep, I do think that is right. using the conditional in the filename like that w...
Oh, wow. Ok. I figured I was way off. It's going into a PR anyway for discussion, so if it needs to be fixed, it can be. But I wanted to at least start from something useful.
tulip sleet
@onyx hinge I tagged you with a testing request here: https://github.com/adafruit/circuitpython/pull/4862#issuecomment-856291286 which you may miss because it's down page. Thanks. This is all pretty mysterious but maybe we can get it to work on all the flaky SAMx chip/boards we have. Tnx.
lone axle
and both will have CODE_OF_CONDUCT.md inside of the if part so that way whichever does get created will be created with that name.
idle owl
and both will have `CODE_OF_CONDUCT.md` inside of the if part so that way whiche...
That's what I was going for.
The files will be different.
lone axle
I hadn't run accross the way that it's using the conditionals in the filenames before diving into cookie-cutter it confused me for a bit but eventually I figured out (I think at least) it works the exact same way as it would if it were inside of a file instead of part of the filename.
idle owl
I mimicked what was there and guessed at what it should look like.
idle owl
@slender iron We still haven't moved adabot to main - there are a lot of moving parts there. We're working on it. That PR should be merging into master still.
idle owl
I'll merge it.
tough flax
Hi Folks,
@tulip sleet and I were testing the Dynamic USB descriptor stuff & we found a problem in the HID Report structure. Basically, CP expects a 1:1 mapping of HID device to Report ID. Some HID devices (most importantly for me, Head Trackers and Eye Trackers) define multiple IN, OUT, and FEATURE report types on a single device.
Has this been discussed since we went over it last week? Or should I create an issue? The consensus seemed to be this should be fixed before 7.x since it changed the API for the Dynamic USB stuff.
tulip sleet
Hi Folks,
<@329766224093249548> and I were testing the Dynamic USB descriptor st...
I plan to fix this before 7.0.0 final. I will make an issue.
atomic summit
Does anyone know what would happen if the there was more than 1 RGB LED (WS or APA) connected to the #define MICROPY_HW_NEOPIXEL (&pin_GPIO18) ? @slender iron will CP just use and display on eh first one?
Like if I had 2 addressable LEDs in the chain, does it just assume it's one and only set the colour on index 0?
That's what I assume it would do, but I wanted to check first
blissful pollen
Does anyone know what would happen if the there was more than 1 RGB LED (WS or A...
I don't have a way to test it but was looking and the code sets 3 bytes and writes them for the status, so I'd assume you are correct it should only write the first LED in the chain
atomic summit
I don't have a way to test it but was looking and the code sets 3 bytes and writ...
Thanks Mark - I was going to check the code later today - but thought I might ask here to avoid digging.
blissful pollen
Thanks Mark - I was going to check the code later today - but thought I might as...
If you're looking it is in supervisor/shared/sattus_leds.c in new_status_color()
I was curious so figured I'd learn where it is too
slender iron
Does anyone know what would happen if the there was more than 1 RGB LED (WS or A...
ya, I think it'll only send one pixels worth of data
atomic summit
ya, I think it'll only send one pixels worth of data
Thanks for the confirmation!
gloomy shuttle
<@!517804297401008138> Disconnected you from audio chat so nobody shows up and s...
Thank you! I didnโt realize I didnโt disconnect before heading out for the day.
idle owl
This PR will pass eventually. When I follow the guide I wrote with the list of steps to do before a PR. ๐คฆ
idle owl
@still zephyr https://github.com/adafruit/Adafruit_CircuitPython_seesaw/pull/73 if you have a moment.
ember iris
<@!607375730610798634> https://github.com/adafruit/Adafruit_CircuitPython_seesaw...
Ah the downsides of not having the hardware to test and just sanity check. Thanks for replying to my questions though!
idle owl
Yeah for sure! Thanks for checking into it!
@tulip sleet If I turn off CIRCUITPY, does it show up if I enter safe mode? Was that part of the deal?
idle owl
Ok thanks.
still zephyr
<@!607375730610798634> https://github.com/adafruit/Adafruit_CircuitPython_seesaw...
I arrived to late to the party, : @ember iris Thanks.
idle owl
I arrived to late to the party, : <@!459119350851567626> Thanks.
No worries! ๐
onyx hinge
@ionic elk is there any easy way to fully disable sleep modes on esp32-s2? (modifying source is fine)
tulip sleet
sleep is not used unless you ask for it
onyx hinge
context: my current project (cameras, esp32-s2 and circuitpython) is triggering watchdog resets in a way that makes no sense and I found https://github.com/espressif/esp32-camera/issues/143
time.sleep() doesn't use low power modes?
idle owl
@tulip sleet Am I seeing right it's two short yellow and one "long" yellow blink on start-up for CP boards? I checked three and that's what it feels like is happening, but I want to verify. Then three short blinks intermittently while in safe mode.
tulip sleet
it uses ``vTaskDelay()`. The "light sleep" and "deep sleep" are special modes that shut things down. We don't use light sleep
onyx hinge
what's going on is weird, there's something I don't understand, and it's making me frustrated.
First you give the camera a square wave on its external clock input. Then, you configure it over I2C. After a few I2C configuration steps, it starts pumping out signals on about 11 pins that are connected to the microcontroller. And then the esp32-s2 watchdog-resets.
tulip sleet
unless you are calling exit_and_deep_sleep() it should not go into a sleep mode
onyx hinge
if the clock is being driven by external hardware, so that the pins are still being driven, it keeps resetting before even booting circuitpython again
(if it's driven by pwm, the pwm is reset when the watchdog bites and can successfully boot CP, or you can power it off and back on)
tulip sleet
hmm, the watchdog may be due to FreeRTOS getting starved
onyx hinge
so I can say some things about that.. I never configure the pins / set up the "parallel capture" peripheral, so they are unconfigured inputs and intuition says they can't affect the execution of code
tulip sleet
are you using any of the "strapping pins" by accident?
onyx hinge
no, they had that bug in the Kaluga v1.2 and fixed it in 1.3.
the pins that end up being driven by the camera are ``` { MP_ROM_QSTR(MP_QSTR_CAMERA_PCLK), MP_ROM_PTR(&pin_GPIO33) },
{ MP_ROM_QSTR(MP_QSTR_CAMERA_HREF), MP_ROM_PTR(&pin_GPIO3) },
{ MP_ROM_QSTR(MP_QSTR_CAMERA_VSYNC), MP_ROM_PTR(&pin_GPIO2) },
{ MP_ROM_QSTR(MP_QSTR_CAMERA_D2), MP_ROM_PTR(&pin_GPIO36) },
{ MP_ROM_QSTR(MP_QSTR_CAMERA_D3), MP_ROM_PTR(&pin_GPIO37) },
{ MP_ROM_QSTR(MP_QSTR_CAMERA_D4), MP_ROM_PTR(&pin_GPIO41) },
{ MP_ROM_QSTR(MP_QSTR_CAMERA_D5), MP_ROM_PTR(&pin_GPIO42) },
{ MP_ROM_QSTR(MP_QSTR_CAMERA_D6), MP_ROM_PTR(&pin_GPIO39) },
{ MP_ROM_QSTR(MP_QSTR_CAMERA_D7), MP_ROM_PTR(&pin_GPIO40) },
{ MP_ROM_QSTR(MP_QSTR_CAMERA_D8), MP_ROM_PTR(&pin_GPIO21) },
{ MP_ROM_QSTR(MP_QSTR_CAMERA_D9), MP_ROM_PTR(&pin_GPIO38) },
i guess you could try disconnecting the pins one by one and see if one is triggering something bad
onyx hinge
here's my terminal log https://gist.github.com/jepler/2f30ca0337342956f99189597d7da152 and I believe that 0x4004c6bf is trying to read flash
mortal kernel
Hey danh, just did some testing on #4867 with my stable of RP2040 boards, you can see the results above. I'm motivated to get CircuitPython nvm working on QtPY boards, so I'm ready to devote some time to getting this resolved. Any suggestions / ideas?
onyx hinge
so the remaining observations are (A) this has never happened to me in the esp-provided demos and (B) I had something an awful lot like this on samd51, where it would "just flake right out in poorly defined ways like memory corruption" in a similar circumstance
(and B.1, the arduino samd51 camera example doesn't have any problem like that either)
ember iris
<@!329766224093249548> Am I seeing right it's two short yellow and one "long" ye...
Would an opencv decoder be of any use in these cases?
idle owl
Would an opencv decoder be of any use in these cases?
I don't even know what that is ๐
onyx hinge
which leaves me very tempted to draw this "it's some general circuitpython problem" line between the two behaviors but that's just implausible
ember iris
I don't even know what that is ๐
I was thinking of writing a program so you could point your webcam at a board as it blinks and it would try to decode the color blinking into the meaning. Opencv is just the program I was thinking of using for it
idle owl
I was thinking of writing a program so you could point your webcam at a board as...
Oh! lol. I don't think all of that is necessary. It'll be in the CircuitPython code somewhere.
tulip sleet
Hey danh, just did some testing on #4867 with my stable of RP2040 boards, you ca...
i would suspect the actual read/write code for flash. @slender iron is much more familiar with that, but is going on vacation soon. So there's something different about the flash chip access for the original Pico vs the other chips. I'd check the board-specific files, and the nvm.toml files
tulip sleet
which leaves me very tempted to draw this "it's some general circuitpython probl...
are you actually trying to read the data, or it's just arriving on the pins?
onyx hinge
@tulip sleet no, I do NOTHING with those pins. I do not configure them. I do not read from them.
tulip sleet
so if you disconnect one at a time, or all of them, what happens?
onyx hinge
All I do is (A) start the external clock with PWMIO and (B) send I2C configuration commands until the clock starts driving its pins
that's easier said than done, it's just a 2x9 header
tulip sleet
maybe the PWMIO is an issue? if you omit the i2c config, it does not happen?
onyx hinge
that's right. I can eliminate PWMIO by driving it with an external clock instead -- the one created by the espressif demo, say.
tulip sleet
are you sure the i2c config is finishing? maybe it's getting stuck in the I2C config and timing out inside there?
jaunty juniper
<@!329766224093249548> Am I seeing right it's two short yellow and one "long" ye...
the long one might be a side effect, like not turning it off right at the end of the blink, but when user code starts
idle owl
the long one might be a side effect, like not turning it off right at the end of...
Hmm. Ok, I need to find it in the CP code then.
ember iris
Oh! lol. I don't think all of that is necessary. It'll be in the CircuitPython c...
Sounds good, I wasn't sure if it would be a generally useful utility to have, or if many users asked about their decodings. I figured it a webcam decoder would be useful I could give it a shot before I start on other code
tulip sleet
<@!329766224093249548> Am I seeing right it's two short yellow and one "long" ye...
i see something like that. I think the current pattern is confusing. I think it's blinking during the 1sec press-here-for-safe-mode time, but there are only two blinks there or something. I would prefer it go back to a steady yellow during that time.
onyx hinge
if I connect just vcc/gnd/sda/scl/xclk, then it gets approximately 240 lines through my initialization code and stops with some kind of i2c I/O error
tulip sleet
or maybe it is actually on for 1sec on yellow -- I am just going by what I heard, not what I observed
onyx hinge
but who knows, maybe introducing the 6inch jumper wires has affected something
tulip sleet
maybe take a break for a bit and then come back to this. is the i2c initialzation expecting one of the unconnected pins to be high or anything?
idle owl
i see something like that. I think the current pattern is confusing. I think it'...
I'm running latest and it seems to blink, on any board, twice fast, one slow and then start up. Checked three.
tulip sleet
i'd have to look at the code. will do that
onyx hinge
no, my python code absolutely doesn't touch the pins that are being driven by the module, except sda/scl
tulip sleet
I'm running latest and it seems to blink, on any board, twice fast, one slow and...
yeah, it is doing a complicated blink pattern during the 1000msec wait. I would propose this be simply a solid yellow instead. SAFE_MODE color is YELLOW in the code below:
while (diff < 1000) {
#ifdef CIRCUITPY_STATUS_LED
// Blink on for 100, off for 100, on for 100, off for 100 and on for 200
bool led_on = diff > 100 && diff / 100 != 2 && diff / 100 != 4;
if (led_on) {
new_status_color(SAFE_MODE);
} else {
new_status_color(BLACK);
}
#endif
I'm writing up a guide section on entering safe mode. That's what brought all of this up.
tulip sleet
the while (diff < 1000) is counting down 1000 msecs
idle owl
If we're going to make that change, ASAP would be good.
Solid might be better, yeah.
"While it's yellow, hit reset". "While it's fast blinking twice then slow blinking once, hit reset"....
tulip sleet
@slender iron how would you feel about making the safe-mode delay be just solid yellow? I didn't know your motivation for the fancy blink pattern during the 1000msecs (see above)
@onyx hinge I am at a loss to explain, but I am suspicious of an I2C error during setup if connetcing the other pins should notmatter. But it seems it does
idle owl
@slender iron Unless you have a really good reason for the blinking at startup, I agree with Dan. I'm typing out this explanation of entering safe mode and it feels unnecessarily complicated.
onyx hinge
@tulip sleet D4 is GPIO41 is "MTDI", which is part of JTAG/SWD
slender iron
I'm in a meeting
onyx hinge
OK, well, that's explained
manic glacierBOT
slender iron
<@!252717193496756235> Unless you have a really good reason for the blinking at ...
@tulip sleet I am simplifying it with these BLE changes to be a consistent blink. I don't want solid though because that then relies on color alone. With the BLE changes it'll be a slow blink for 1 second for safe mode and then a faster blink for 1 second to erase the BLE bonding info and start over
the second BLE phase will be solid when it is successfully entered. We could do the same thing for safe mode too
tulip sleet
makes sense. The current irregular blink I think is hard to describe, and seems like two separate thing happening in sequence. Maybe a 100ms on/ 100ms off would be clearer, and an even faster bink for BLE bonding erase.
idle owl
A consistent blink would at least be easier to explain.
tulip sleet
anything that's only one or two blinks is hard to recognize and discriminate between that and something else
onyx hinge
double checking the alt-functions of the parallel capture pins on samd51 doesn't turn up anything so interesting, so it's not an explanation there. oh well.
idle owl
@slender iron So will it be a 100ms blink? Is that the plan? Or should I tell Limor I need to hold off on explaining how to enter safe mode for now because it's going to change? Or document the way it is now and make a note to go back and change it when you submit changes?
slender iron
ya, just wait until it is merged in
idle owl
Ok.
idle owl
like the cake.
slender iron
yup. I think for a learn thing you could document it as it is in 6.x
and say it'll change in 7.x
tulip sleet
it is solid yellow for 700ms in 6.x
idle owl
Ok. I guess that works.
idle owl
Wait, we didn't fix the NeoPixel power pin thing in 6.x did we.
So I need to test on something other than RP2040.
Erm..... I was supposed to be documenting how to get an RP2040 into safe mode.
@slender iron I can't document how it works in 6.x for the RP2040. It apparently doesn't.
slender iron
right. we did have that bug didn't we
idle owl
Indeed
slender iron
where are you documenting this? I don't have context on why you are doing it now
idle owl
Limor asked me to. It's going in the template for Installing CircuitPython on RP2040.
idle owl
where are you documenting this? I don't have context on why you are doing it now
If you think it should wait, I can let Limor know. But that's the context.
slender iron
safe mode should work in 6.3.0
idle owl
I just loaded it onto an Itsy and no yellow.
slender iron
try doing the click anyway
idle owl
Ok
idle owl
So document it for now as "click reset and then click it again w/in 700ms" ?
It worked.
slender iron
so it's not all RP2040 boards, just the ones that require power to the led
idle owl
Ok. I'll document both then.
slender iron
up. I think of it as a slow second tap
idle owl
Ok.
idle owl
Two out of three of our RP2040 boards require NeoPixel power.
ยฏ_(ใ)_/ยฏ
Well two out of three I have. The Trinkey is on the way.
slender iron
ya, that'll be fixed in 7.x with the blink revamp
safe mode should still work. it'll just be harder to know when to click
idle owl
I made it work on the first try with the Itsy. So I think it's fine.
slender iron
๐
jaunty juniper
yeah, it does not run any user code on boot, and disables auto-reload
idle owl
Thanks
jaunty juniper
you can still use the REPL and import code from there, but you probably don't need to mention that
ornate breach
Thank you @gilded cradle ๐
idle owl
Safe mode in CircuitPython does not run any user code on startup, and disables auto-reload. This means a few things. First, safe mode bypasses any code in boot.py (where you can set CIRCUITPY read-only or turn it off completely). Second, it does not run the code in code.py. And finally, it does not automatically soft-reload when data is written to the CIRCUITPY drive.
Therefore, whatever you may have done to put your board in a non-interactive state, safe mode gives you the opportunity to correct it without losing all of the data on the CIRCUITPY drive.```That's what I have so far.
There's some bold text in there. But that's the idea.
ornate breach
Iโll have an updated version of the board copy this evening so the description matches the board update. ๐
idle owl
@tulip sleet Am I understanding correctly that the safe mode changes that are happening in 7.x are changes to the LED behavior only? The time to press reset again remains the same?
tulip sleet
time is increasing from 700ms to 1000ms, on my request
tough flax
Does anyone know of a way in CP to check whether a pin has a pull DOWN resistor on a given pin? I suppose I'm good with a list of which boards/pins have them and I can handle it, but is there such a source of metainfo about pins at runtime?
slender iron
an external resistor?
tulip sleet
Does anyone know of a way in CP to check whether a pin has a pull DOWN resistor ...
do you mean whether it is capable of having a pull-down or not, or do you mean testing whether there is an external resistor?
you could turn on the internal pull up resistor on the pin and see what the state is. The internal pull-up resistors vary per chip. On some chips they are like 20k, on RP2040 they are 50-80k. If your external pulldown is stronger, then it will override the internal pullup
ionic elk
@onyx hinge fully disable? I mean you could just get rid of the whole module, remove alarm altogether in the macro list? Why do you ask?
onyx hinge
@ionic elk oh that turned out to be a red herring, you can read back to the conversation I had with dan if ya care
the real culprit: sending waveforms into the JTAG debug pins
ionic elk
@onyx hinge yeah I'll check it out
onyx hinge
sleep mode is blameless
ionic elk
@tulip sleet I am still getting sort of surprisingly bad power performance on NRF. My Circuitplayground Bluefruit can't seem to go below 6mA in power consumption even in deepsleep
how bad should I expect the power cost to be when connecting the PPK2 to the Bluefruit via the holes? I'm using tightly screwed M3 screws and washers to hold the supply pins directly onto the exposed copper, would that lower power efficiency by this much? I figured it couldn't be that much different than the tin pins in the battery header...
But I also recall the NRF feather seemed similarly poor
looking at my notes, the feather was a bit better, actually, but still high - 3.5mA in deep sleep
ugh on the circuitplayground my ppk says that light sleep is actually better than deep sleep. That might even be worth an issue.
tulip sleet
you are probably measuring the consumption of the power LED, the regulator, and some other stuff that is always powered, like the LIS3DH, the IR receiver, etc. The CPB is not a good candidate for low-powering. The Feather Sense has some things you can turn off. Also you should power the Feather through the battery terminals
you might try an nRF dev board, powered directly from the PPK, not throuhg USB
ionic elk
wait I thought that the CPB was the one I was supposed to get? I had a note for it and everything
tulip sleet
check the schematic about what's on the 5V line
and the 3.3V line
i don't remember why you were asked to get the CPB. We don't really have an nRF board that was designed for low power.
ionic elk
Yeah, you're right, there probably is a lot of stuff leaking.
tulip sleet
do you have an nRF dev board?
ionic elk
No, I don't have any non-adafruit NRF stuff
tulip sleet
were you powering the feather through the battery terminal?
ionic elk
VBAT,
so basically?
minus the transistor
or do you mean the CPB? That I'm powering with GND and VCC, which is also basically the same as the battery terminal minus a protection diode
tulip sleet
yes, VBAT. I have a JST cable that has Dupont headers on one side and the JST battery-style connector on the other, and just connect it to the PPK that way
ionic elk
Yeah I've just been using VBAT. I like it to be easy to plug/unplug - do you let the PPK stay plugged in when reprogramming the board with USB?
I've been unplugging it every time
tulip sleet
I use this: https://www.adafruit.com/product/3064. I honestly don't remember if I was careful to turn it off each time to avoid backpowering the PPK. Maybe the PPK has a diode or some other backpowering protection. It might say in the docs
tulip sleet
if you get that switch, note that it is ON when the button is up. Kind of confusing. I ultimately put a label on it to remember that.
idle owl
That is weird about that cable. I remember you saying that. I still haven't had to use it yet, so I haven't run into it, but guarantee, I've read it at least twice from you, and I will still forget and confuse myself.
tulip sleet
idle owl
Love it.
blissful pollen
As I just got myself a PPK the two wires leaving the button go to the VBAT/GND or feather JST input generally? before I accidentally fry anything!
ionic elk
@blissful pollen if you look at the schematics for most boards, VBAT and GND are directly connected to the JST pins, maybe sans a protection diode or something
blissful pollen
Thanks, was just going to say make sure I don't swap + and - !
slender iron
@tulip sleet you around to chat bonding?
tulip sleet
ok, sure, one minute
slender iron
kk, brb myself
tulip sleet
call me when u r back
sour yew
What's the status/plan for (expanding?) the dynamic HID descriptors in circuit python? My understanding is, with that, you can create descriptors/devices on-the-fly, which sounds really useful, but it appears that only one at a time is currently supported. Is that right?
slender iron
What's the status/plan for (expanding?) the dynamic HID descriptors in circuit p...
this probably gives the best overview of what we could add from what we have: https://github.com/adafruit/circuitpython/issues/4868
tulip sleet
What's the status/plan for (expanding?) the dynamic HID descriptors in circuit p...
you can have multiple devices in one USB HID interface. RIght now there is not support for multiple interfaces (if you wanted a boot keyboard, etc.).
sour yew
In my current implementation I'm using the boot keyboard and absolute mouse from the updated Arduino repository: https://github.com/NicoHood/HID and it seems to be working so far.
I would really like to be able to have two unique devices interfaces (to connect to two hosts at once), but I think that's a stretch to far for the short term. ๐ All in circuitpython, of course.
I'm not sure I fully grasp WebUSB yet, but from what I've seen, it seems like, if you had a device (STM32F4, teensy, etc), that had two USB device interfaces, that you could make a device that plugged into a PC, and was able to emulate arbitrary HID devices on the other interface. It that completely off base?
solar whale
@onyx hinge FYI -- did you see that there have been some recent (15 days ago) updates to the esp32s2 camera-driver code used for the kaluga? https://github.com/espressif/esp32-camera/tree/fa5020f5c75df4f4aeaaf7d7a7ddceb84969c604 looks like OV7670 was added!
tulip sleet
In my current implementation I'm using the boot keyboard and absolute mouse from...
Just to be clear, when I said two interfaces, I meant USB descriptor interfaces, all connected to the same host. We don't have support for multiple USB PHY interfaces.
onyx hinge
@solar whale yeah, I did see that
solar whale
@onyx hinge Did you get the sdk demo to work on the OV2640? I get a very garbled picture on my screen.
onyx hinge
@solar whale I've seen various misbehaviors, but I have gotten a relatively stable display with OV2640 on a v1.3 kaluga.
solar whale
do you hav the st7899 or ili9341 screen?
onyx hinge
one of each.
solar whale
how can you tell?
onyx hinge
you can't, you try menuconfigging for both of them until one works
even when st7899 works, in their demo the screen is rotated wrong
it's .. not great
solar whale
thats what I did -- ili9341 give me the garbled screen st7789 gives me nothing ... did you have to set anything in else in menuconfig?
onyx hinge
with a v1.3 kaluga, no, that's the only setting I've messed with
I tried various settings trying to get a v1.2 to work but never got anything
solar whale
OK thanks -- I'll keep at it.
onyx hinge
I .. feel like .. the hardware or maybe the software is not super reliable. and it worries me we won't be able to provide a solid circuitpython implementation either
solar whale
Hopefully "it's only software" ๐
plucky briar
Hello everyone! trying to teach myself Python and more importantly CircuitPython. I'm trying to make a wire cable tester that i turn on outputs and then make sure the correct inputs are turning on. How do i compare a list of input values (True or False) to a list for each wire number that needs to be True?
say i have wire 1 output turned on, then i need to look at all my inputs and make sure input 1 and 5 are on and the rest are off. wire 2 output turned on, then i need to see inputs 8, 10 and 12 on and the rest off. I'm sure there is an easy way of doing this but I'm not sure how. I have 4 MCP23017 chips that I'm using, 2 as inputs and 2 as outputs.
I'm able to turn on each individual wire and see what inputs are turned on but i would like to have it automatically compare them and give me a thumbs up or tell me what wire is not correct.
orchid basinBOT
orchid basinBOT
idle owl
Hello everyone! trying to teach myself Python and more importantly CircuitPython...
Are you expecting this to be done in hardware, or are you wanting to compare things in software?
@plucky briar Also you might get better assistance in #help-with-circuitpython . This channel is more for development discussion. We discourage cross-posting, but in this case it's ok. I would suggest posting your question there.
slender iron
@candid sun or anyone with arduino rp2040: https://forums.adafruit.com/viewtopic.php?f=60&t=180126
candid sun
<@!347203168121126912> or anyone with arduino rp2040: https://forums.adafruit.co...
I havenโt tried that but I can take a look
slender iron
rp2040 should autodetect the size using the flash's jedec info but I wonder if its wrong for the flash
let me know if you see the same thing
(I'm just about to head to the dentist but I'll be back on later)
I think the rp2040 code defaults to 1mb if the jedec value is out of range
candid sun
I think the rp2040 code defaults to 1mb if the jedec value is out of range
yeah it is showing at 1MB for me too
ornate breach
Hey @slender iron, does the SPI flash for the rp2040 have to be qspi specifically or can standard SPI Flash work as long as itโs defined?
I can get my boards to enumerate over usb but I havenโt attempted to make a circuitpython build yet
Enumeration leads me to believe that standard SPI flash is fine.. but I havenโt completely deciphered any different thinking
slender iron
You really do want qspi for speed since code is stored on the flash
ornate breach
Okay, Iโll have to do that on my next revision.
lime trellis
<@!492049388152029185> <@!329766224093249548> off this week but we can definitel...
Thank you for the AWESOME flow chart on the github issue. I'm trying to get a summer student to work on implementing this. Will keep you updated.
orchid basinBOT
solar whale
@onyx hinge Have you gotten the ILI9341 board (ESP-LyraP-LCD32 V1.2) to work with the Kaluga under CP? Just curious? Not asking you to get it working.
trim elm
If anyone here is good with memory stuff it'd be amazing if you could take a look at my question in #help-with-circuitpython
onyx hinge
@solar whale yes I think that's the one I got Working in CP. Try the code in that issue I replied to
solar whale
@onyx hinge Thanks -- that works!
slender iron
@candid sun I started an issue for it here: https://github.com/adafruit/circuitpython/issues/4874
tough flax
you could turn on the internal pull up resistor on the pin and see what the stat...
Sorry, I posted this and my wife sprung a surprise trip to a rays game on me :-). Ok Iโm not that sorry. Iโm aware that most pins have internal pull-up resistors that you can turn on. However I also know many (samd21/51 at least) also have internal pull down resistors. Iโm just wondering if thereโs a way to tell without going to the data sheet on each microcontroller.
tulip sleet
Sorry, I posted this and my wife sprung a surprise trip to a rays game on me :-)...
As far as I know, all the chips we support have both internal pull-ups and pull-downs. The main difference on GPIO's is that several don't inherently support open-collector outputs. Instead you fake it switching between input and output depending on the level. The nRF chips have genuine OC outputs that don't require that.
tough flax
As far as I know, all the chips we support have both internal pull-ups and pull-...
I really hope you're right on that - it will simplify things tremendously. We have a number of devices we make that we are refactoring to work with Microsoft's XAC standard. That is that switches are provided 3.3V and send 3.3V back when "pressed". In other words they use pulldowns on the XAC. If we switch to their approach, we'll either have to add pulldowns on every board or have the pins support it. Since most of these use a Feather, Itsy, or QT sized board, I want to make sure that nRF, SAMD, and RP2040 based boards can all be used
tulip sleet
I really hope you're right on that - it will simplify things tremendously. We h...
nRF, SAMD, RP2040, ESP32-S2 and STM all support both pullups and pulldowns. I have never heard of a chip that doesn't have pulldowns
i looked at the datasheets to double-check, but really, I wouldn't have assumed otherwise
tough flax
I know that the ESP8266 only had pulldown on one pin (pin 16) or something like that
tulip sleet
yet another reason to drop support for such an idiosyncratic chip
tough flax
Yes yes, I am happy with your answer, just explaining why I asked ๐
tulip sleet
np ๐ that did jog my memory, but that chip had such feeble I/O. It was built for wifi lightbulbs
cobalt grail
Not really CP related, by the ATMega only supports pullup. Jogged my memory. ๐
tulip sleet
Not really CP related, by the ATMega only supports pullup. Jogged my memory. ๐
@tough flax OK, AVR's don't have internal pulldowns. I am both sorry I didn't know that and glad I never had to deal with them not being there. No CPy on AVR.
slender iron
@tulip sleet here is the keyscanner code from @tall vigil: https://github.com/keyboardio/avr_keyscanner/blob/master/firmware/debounce-split-counters.h
as a reference
orchid basinBOT
tall vigil
<@!329766224093249548> here is the keyscanner code from <@!299335186246795266>: ...
I'm happy to answer questions or braindump about what we've learned about keyscanners and debouncing over the years.
hollow token
@tulip sleet how does that work with more than one row?
oh, just saw the commit messages, it doesn't (yet (tm))
tulip sleet
right, not yet
orchid basinBOT
tulip sleet
@idle owl @slender iron @onyx hinge FYI the "latest" readthedocs build has been failing for a while. Example: https://readthedocs.org/projects/circuitpython/builds/13973349/. I'm not sure what's wrong at first glance.
idle owl
<@!330227457296957440> <@252717193496756235> <@!426200547024961539> FYI the "lat...
Email me, please. I'll take a look at it in a bit.
onyx hinge
@tulip sleet seems like maybe it has to do with virtualenv, except these are all super old bugs? https://github.com/pypa/virtualenv/issues/737 https://github.com/pypa/pipenv/issues/3824
tulip sleet
Not sure. I note they are buildling with 3.7.
onyx hinge
that code was recently added fwiw
it has to do with making the stubs pypi'able, I think
tulip sleet
we added code recently? I guess we need an issue on that.
onyx hinge
on examination, I am not confident that is right anyway
it builds a whl file with entries like usr/local/lib/python3.9/dist-packages/_bleio/__init__.pyi while an (I assume working) lxml-stubs package from pypi has lxml-stubs/__init__.pyi
@lone axle do you have the time to chime in about this?
lone axle
Yes, but stepping away for just a few minutes then back
lone axle
it builds a whl file with entries like `usr/local/lib/python3.9/dist-packages/_b...
back now. to make sure I understand: you are running a pip install and ending up with the stubs files inside of dist-packages not site-packages ?
onyx hinge
<@!330227457296957440> <@252717193496756235> <@!426200547024961539> FYI the "lat...
@lone axle We are looking at the stubs thing due to the build failures of readthedocs, see replied-to message
I was comparing what goes in a circuitpython-stubs package (built by bdist_wheel) as compared to one from pypi and I noticed that the paths are different .. it is more like a "full path" in circuitpython-stubs but a partial path in lxml-stubs
but I think the real thing to resolve first is that our readthedocs isn't building now, and it's likely that it stopped working after we merved #4844
lone axle
the code inside of build_data_files_list() is attempting to make a list of tuple pairs that represent the stubs files and where they should end up when installed like this:
[
("lib/python3.8/site-packages/digitalio/", ["circuitpython-stubs/digitalio/__init__.pyi"])
# ...
]
the part that seems to be failing in the readthedocs message is the code that is meant to programatically find the location of the site-packages directory for the environment that it's being run inside of.
weird that it seems the function is not existing in that environment. this docs page shows it was added in 3.2:
https://docs.python.org/3/library/site.html
the quickest temporary fix is probably to just remove that code from setup.py. The rest of the actions are not set up yet to deploy to PyPi anyhow so removing this back out won't have any other impact at the moment.
For a longer term fix more generally we need to figure out a way to make the pip install place all of the stub packages into the root of site-packages/ the more "standard" (I think at least) way of doing setup.py pip installation makes it so all of those stubs end up inside of site-packages/circuitpython-stubs/ which means that the IDE cannot find them unless the imports are changed to be like from circuitpython import digitalio but that would make the code incompatible with microcontrollers.
Is there a guide or any documentation around building the docs in the circuitpython repo? I tried the same command inside of docs that I use for libraries but that is not working in the main repo. I'm curious if the same error from readthedocs happens locally or not.
tulip sleet
Is there a guide or any documentation around building the docs in the circuitpyt...
I just cd circuitpython; make html.
that bug did not happen for me locally; I build locally to get errors out of my docstrings before committing.
lone axle
nice, I really should look through the make file to learn all of the available commands.
I was missing some other required libraries, after getting them installed I think the build still failed for me though:
I don't think I understand what the specific issue is that caused it to fail though. Perhaps I am missing something required from my PATH.
ah, need to apt install librsvg2-bin possibly
yep, I got a successful build locally as well after that.
My best guess is perhaps the container that it's running inside of at readthedocs had that function removed possibly to prevent code from adding things to site-packages. I don't really have any ideas of what we could do differently in setup.py to end up with the stubs installed in they way they need to be for the IDE to find them. It seems a little odd to me that readthedocs is even trying to execute the setup.py I'm not sure why it would need to do that for the docs build process.
onyx hinge
The failure is for sure something specific to readthedocs...
onyx hinge
mypy is kinda neat but dang its mypy -c '...' is not fast for one-offs.
idle owl
@onyx hinge @tulip sleet @lone axle Did this get sorted? Or is it still up in the air. Someone should file an issue on CP and tag the four of us if it's still not resolved. This discussion is a lot of information to not have documented.
onyx hinge
@idle owl I am writing it up from my POV now
idle owl
Thank you!
gloomy shuttle
@tulip sleet I made those changes you requested on the rotary encoder work for the adafruit_circuitpython_seesaw library. https://github.com/adafruit/Adafruit_CircuitPython_seesaw/pull/70
idle owl
@gloomy shuttle Does that change the user-facing behavior of it? e.g. do I have to update guides if this is merged?
gloomy shuttle
<@!517804297401008138> Does that change the user-facing behavior of it? e.g. do ...
No, the changes are in seesaw.py which isn't in any of the examples. It does change the direction so clockwise is positive and counter-clockwise is negative to match the standard rotaryio.
idle owl
Hmm. OK. It might make my descriptions of what is happening when rotating backwards. ๐
I guess I'll have to test it when it's in.
tulip sleet
<@!329766224093249548> I made those changes you requested on the rotary encoder ...
I didn't approve this yet because Limor didn't necessarily want to change it. I think she would be more likely to change the firmware for future products. I will ping her in the PR to see whether to go ahead or just close it.
gloomy shuttle
I didn't approve this yet because Limor didn't necessarily want to change it. I ...
That makes sense and seems reasonable.
I'll go ahead and close it for now