Blergh.
#circuitpython-dev
idle owl
onyx hinge
From https://github.com/raspberrypi/pico-sdk
! [rejected] 1.3.0 -> 1.3.0 (would clobber existing tag)
jepler@bert:~/src/circuitpython/ports/raspberrypi/sdk$ git tag -d 1.3.0
Deleted tag '1.3.0' (was ebc601f)
jepler@bert:~/src/circuitpython/ports/raspberrypi/sdk$ git fetch origin --tags
From https://github.com/raspberrypi/pico-sdk
* [new tag] 1.3.0 -> 1.3.0
``` for me, this fixed it(manually deleting the tag that they replaced/changed/updated)
onyx hinge
good luck!
idle owl
Also, git status is showing all the submodules updates. Now I have this: modified: ports/broadcom/firmware (untracked content) modified: ports/stm/st_driver/CMSIS_5 (untracked content) modified: tools/Tecate-bitmap-fonts (untracked content) modified: tools/adabot (untracked content) modified: tools/huffman (untracked content) modified: tools/uf2 (untracked content)
onyx hinge
I almost always ignore 'untracked content' messages
idle owl
Hm ok. I always use git add . so I usually don't ignore them, heh. Probably spend less time adding specific files than I do going through all of those directories and figuring out what changes to delete, simply so I can do git add ..
But half the time I do the longer process.
Ok, more than half.
native sandal
Hey @lone axle , thanks for getting back to me. Yes, mypy --disallow-untyped-defs produced those warnings, that's how I approached finding what needed annotation. The problem I'm having is that I can't seem to properly import WIZNET5K so that it's known to mypy as a type name. Anything I try results in
Name "WIZNET5K" is not defined
I even tried adding from adafruit_wiznet5k import WIZNET5K but I get Module "adafruit_wiznet5k" has no attribute "WIZNET5K".
onyx hinge
I try to always use git-gui and select things one by one
idle owl
```jepler@bert:~/src/circuitpython/ports/raspberrypi/sdk$ git fetch origin --tag...
After doing these commands, the make fetch-submodules runs through the whole list.
lone axle
Hey <@382939733107408897> , thanks for getting back to me. Yes, `mypy --disallow...
are you working on a git branch already? If possible can you commit / push your version to github and I could then check out yours to see if I can replicate.
native sandal
Yes, I posted a link to my repo/branch earlier, but here it is again: https://github.com/JetForMe/Adafruit_CircuitPython_Wiznet5k/blob/rmann/issue-45-missing-type-annotations/adafruit_wiznet5k/adafruit_wiznet5k_socket.py
Lemme know if that's not what you meant
lone axle
it's on the annotations like this one https://github.com/JetForMe/Adafruit_CircuitPython_Wiznet5k/blob/bd2219d3b6361d02cd5cf83ed71175d94a518a88/adafruit_wiznet5k/adafruit_wiznet5k_socket.py#L29 that you get the error?
native sandal
it's on the annotations like this one https://github.com/JetForMe/Adafruit_Circu...
As far as trying to annotate that specific type, yes. But there are other errors related to import, too (note that the line numbers are off here a bit because I've been experimenting):
$ mypy adafruit_wiznet5k/adafruit_wiznet5k_socket.py --disallow-untyped-defs
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:19: error: Module "adafruit_wiznet5k" has no attribute "WIZNET5K"
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:30: error: Name "wiznet5k.WIZNET5K" is not defined
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:123: error: Function is missing a type annotation
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:153: error: Module has no attribute "adafruit_wiznet5k"
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:159: error: Module has no attribute "adafruit_wiznet5k"
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:160: error: Module has no attribute "adafruit_wiznet5k"
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:161: error: Module has no attribute "adafruit_wiznet5k"
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:162: error: Module has no attribute "adafruit_wiznet5k"
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:164: error: Module has no attribute "adafruit_wiznet5k"
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:198: error: Module has no attribute "adafruit_wiznet5k"
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:219: error: Module has no attribute "adafruit_wiznet5k"
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:220: error: Module has no attribute "adafruit_wiznet5k"
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:224: error: Module has no attribute "adafruit_wiznet5k"
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:236: error: Module has no attribute "adafruit_wiznet5k"
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:252: error: Incompatible types in assignment (expression has type "Tuple[int, ...]", variable has type "Optional[str]")
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:252: error: Item "None" of "Optional[str]" has no attribute "split"
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:291: error: Module has no attribute "adafruit_wiznet5k"
Most of those error messages are caused by mypy suddenly not liking the imports when type annotation is applied to the containing method. If you remove the type annotations, then those "Module has no attribute" errors go away
(A couple of those errors are because I don't have the annotation quite right)
Can anyone explain why the original file has import adafruit_wiznet5k as wiznet5k but then references all symbols in that module as wiznet5k.adafruit_wiznet5k.<whatever>?
lone axle
@native sandal I was able to get it to recognize the class by doing the import like this:
from .adafruit_wiznet5k import WIZNET5K
then the annotations like this:
native sandal
Huh. What does the leading . do?
lone axle
I am probably not the best suited to answer accurately... My understanding is it tells it to look "in this directory" for the thing trying to be imported
similar concept to ./ in a terminal.
native sandal
Makes sense
I'm still uncomfortable with the other import
import adafruit_wiznet5k as wiznet5k
lone axle
I don't really get why it would be needed specifically in this case because it seems like wiznet5k.WIZNET5K should have existed to me.
native sandal
Agreed
lone axle
I don't have any insight into the way the import is done in the original code either unfortunately.
idle owl
@lone axle So, wait. Will it continue to follow the same actions run as long as it's running? Or on the next loop does it consider it already_shown and not do anything with it?
lone axle
@native sandal if you are interested in doing so you could refactor a bit to change how constants are imported and other tweaks while you're working on typing. Totally cool if you'd rather not tackle it all at once, or at all as well though.
lone axle
<@382939733107408897> So, wait. Will it continue to follow the same actions run ...
it will continue to follow it as it is running. Only once it sees that the workflow is complete will it add it to the already_shown list and then not sure it on subsequent loops.
native sandal
Well, now that we've got the class importing, I'm running into new issues. I have to make _the_interface type Optional[WIZNET5K], and then I get a ton of:
adafruit_wiznet5k/adafruit_wiznet5k_socket.py:77: error: Item "None" of "Optional[WIZNET5K]" has no attribute "get_host_by_name"
stuff like that will have to be ignored at this point. For the first pass through the libraries the main mypy complaints we want to tackle are the missing types for function / return.
When we get the bulk of the libraries taken care of on the first pass then at some point we'll enable mypy checks in the actions, and at that time choose the right configuration that will affect which things it flags.
bot deleted it. I can see though
native sandal
Why did it get deleted? I just thought I had a stroke
lone axle
My guess is the inclusion of "hell". But not sure.
native sandal
Well, I guess I can make it this: _the_interface: Any = None
idle owl
Yes, that was the issue.
native sandal
Any idea what the types should be on these parameters?
def __exit__(self, exc_type, exc_val, exc_tb):
if self._sock_type == SOCK_STREAM:
self.disconnect()
stamp = time.monotonic()
while self.status == wiznet5k.adafruit_wiznet5k.SNSR_SOCK_FIN_WAIT:
if time.monotonic() - stamp > 1000:
raise RuntimeError("Failed to disconnect socket")
self.close()
stamp = time.monotonic()
while self.status != wiznet5k.adafruit_wiznet5k.SNSR_SOCK_CLOSED:
if time.monotonic() - stamp > 1000:
raise RuntimeError("Failed to close socket")
It may end up being the case that some libraries may need more refactoring in some situations in order to appease all of the mypy checks. Some may end up being ignored or configured less strictly.
I'm not familiar with __exit__ but I would guess that CPython has something similar and our arg types should probably match.
jaunty juniper
type is ignored, exc_val would be an Exception instance, exc_tb is the traceback, obtainable with ex.__traceback__
native sandal
type is ignored, exc_val would be an Exception instance, exc_tb is the traceback...
They may be ignored, but I need to know the type for annotation.
jaunty juniper
at least I think, it matches the arguments of traceback.print_exception
(well you don't need to, annotations are optional)
native sandal
Of course, but for completeness…
jaunty juniper
here is the traceback module:
https://docs.circuitpython.org/en/latest/shared-bindings/traceback/index.html
it's the type of the exception, so Type[BaseException] apparently
native sandal
Ugh, the code plays fast and loose with types internally (e.g. host can be a str or bytes and it branches based on type. Makes annotation hard, even with Union
lone axle
it's okay if you end up sticking to more generic types for the variables that are like that. First pass has exposed one or two other libraries with similar things. Not needing to have it solved 100% now though. We're aiming to get the bulk of the straight up missing annotations filled in before we get into the specifics in some of the more convoluted cases.
idle owl
Tomorrow and Wednesday likely don't work for me. Figuring out what sushi we want for dinner, once that's ordered, I should be able to chat. Does that work?
proven garnet
Sounds good, and I'm definitely referring to sushi as well 🍣 
idle owl
😄 Nice
jaunty juniper
I always take the M6
blissful pollen
Just order all the sushi
idle owl
This place has a LOT of options. That would be... inadvisable.
blissful pollen
Not ordering too much sushi is a skill in itself
idle owl
Indeed.
idle owl
Sounds good, and I'm definitely referring to sushi as well 🍣 <:chefskiss:932492...
Ok, sushi sorted. Are you still around?
proven garnet
Yup yup!
idle owl
Ok, let's vid.
idle owl
@lone axle It occurred to me we could use blinking to differentiate queued and running. Limor liked it. I think I can do that, looking at the code, so I'll try that first, and ping you if I can't get it going. I won't get to it until tomorrow anyway.
manic glacierBOT
manic glacierBOT
timber mango
Hello, I 'm designing a sampler with a Pipico and a neotrellis. As the memory is not big enough on the pico, I plugged an SD card with a SPI BUS, but my pico crashes often when I play samples from the SD card. Should I copy the samples to the internal memory of the pico first? If yes, do you know how? I don't understand how to copy the files with the OScircuit python library. Thanks
stuck elbow
well, there is no command to copy a file, you just open the source file for reading, open the target file for writing, and then in a loop read all bytes from the former and write them to the latter
wraith crow
@timber mango In case it helps, here are the instructions I use to copy files. Copying in binary mode is probably overkill for your application but it won't hurt.
fOrig = open("/sd/filename")
fCopy = open("/filename", "wb")
line = fOrig.readline()
while line != "":
fCopy.write(bytearray(line))
line = fOrig.readline()
fOrig.close()
fCopy.close()
onyx hinge
@tulip sleet do you remember if we investigated gcc 11.2 (arm's toolchain binary released in February)?
manic glacierBOT
manic glacierBOT
slender iron
@tulip sleet have you updated the IDF in your work already?
tulip sleet
<@329766224093249548> have you updated the IDF in your work already?
updated to what version?
slender iron
newer 4.4
tulip sleet
no, was waiting for 4.4.2. I could try 4.4-dev, because it may have an I2C fix for S3 (which did not work for me when manually applied)
is 4.4.2 released? It was not the last time I looked
slender iron
ya 4.4.1 is from april
nope
lots of activity on the branch though
there is an IPC fix that may fix deep sleep on the S3
tulip sleet
At one point I tried master, but it had many different API changes
slender iron
ya, I'm not ready for 5.0
tulip sleet
we could try 4.4-dev for sure
slender iron
kk, I'll update it
tulip sleet
since this is alpha, we can keep moving it up
slender iron
yup yup
I don't see the same I2C fix in the 4.4 branch
will keep on our own patched version
should I rebase it?
timber mango
<@456226577798135808> In case it helps, here are the instructions I use to copy ...
sorry but i got this error
'OSError: [Errno 30] Système de fichier en lecture seule'
Does anyone knows how to change that
stuck elbow
does it say "read-only filesystem"?
tulip sleet
I don't see the same I2C fix in the 4.4 branch
I reverted to the espressif branch. I looked it over at the time, and I decided at the time that it was fine not to include our fixes
slender iron
ok, so we aren't using our branch now?
tulip sleet
right, I brought ours up to date (and used a different branch name)
stuck elbow
Does anyone knows how to change that
the filesystem is read-only by default for the mcu, so that it can be read-write for the computer, you can change it in boot.py: https://learn.adafruit.com/circuitpython-essentials/circuitpython-storage
tulip sleet
i'll check the PR's
slender iron
.gitmodules has us using espressif directly
but its set to the commit from our branch
tulip sleet
I think that may be a commit from upstream.
slender iron
tulip sleet
I don't understand where that commit could come from if .gitmodules is upstream, hmm
slender iron
it must be able to get it anyway
so should we keep that I2C fix?
I can make them match
tulip sleet
the fix is from o-marshmallow, so maybe it is in 4.4-dev
he closed the issue https://github.com/espressif/esp-idf/issues/8770
slender iron
release/v4.4 is the active branch I thought
tulip sleet
release/v4.4 is the active branch I thought
ok, I will ask o-marshmallow about where the v4.4 version of the patch is
and you can correct the .gitmodules
slender iron
you don't want to keep the patch? I was thinking I'd make a new circuitpython8 idf branch
tulip sleet
you don't want to keep the patch? I was thinking I'd make a new circuitpython8 i...
oh, no, I do want to keep the patch. Maybe .gitmodules to should back to adafruit/esp-idf
?
there is no other source of the patch now, apparently
slender iron
@tulip sleet kk, I'll keep it
timber mango
the filesystem is read-only by default for the mcu, so that it can be read-write...
thanks, but does it works for a Pico pi ?
manic glacierBOT
wraith crow
@timber mango The instructions @stuck elbow provided should work fine for the Pi Pico, to change the read-only mode you'll need to access the Pi Pico from your computer using the mounted CIRCUITPY drive, but once you have the Pi Pico connected to your computer there's probably no need to run the python commands to copy the file, just copy the file you want from the SD card to the CIRCUITPY drive on your computer.
minor plume
@tulip sleet FYI, I just submitted an issue (https://github.com/adafruit/Adafruit_CircuitPython_asyncio/issues/23) for an error I'm getting on importing asyncio on a MatrixPortal M4 with 7.3.1. I'm intending to dive deeper and see if I can narrow down what's going on here, but I'm calling it out in case it's a more widespread problem than just that hardware, since I may not have time to work on it before the end of the week.
slender iron
and this version of the idf doesn't start up....
slender iron
@tulip sleet how is esp32 going?
tulip sleet
I have it starting up, and it's now failing when doing something in the heap, I think. Your backtrace tool is extremely helpful. It was boot-looping because there were more pins that should not have been reset than the obvious SPI flash and PSRAM ones. I now have the full list: I just printed out each pin before resetting it, and found the ones that caused crashes.
I was out until mid afternoon due to a long appt
slender iron
nice! glad it was useful
I'm in the IDF weeds because the latest 4.4 doesn't start up
tulip sleet
oy!
tulip sleet
good sleuthing!
slender iron
yup, works after a revert
lone sandal
manic glacierBOT
timber mango
<@456226577798135808> The instructions <@351136009922019331> provided should wor...
@wraith crow , I managed to authorize writing from the pi pico but now I'm blocked. How do I reset the write permission from the pipico from my computer. Thanks a lot.
timber mango
Hello, I' m making a drummachine with a picopi and a neotrellis. The drummachine works with samples (.wave files). As the memory on the pico is very short I decided to plug an SDcard to store sample, but when I play a sample I often got OSError: [Errno 5] Erreur d'entrée/sortie.
Do you think, playing sample from an SD card is a good idea (stable way) ?
Do you think the problem is with my SD card reader ?
jaunty juniper
Hello, I' m making a drummachine with a picopi and a neotrellis. The drummachin...
note that questions about using Circuitpython are better in #help-with-circuitpython , this channel is for the development of Circuitpython
timber mango
note that questions about using Circuitpython are better in <#537365702651150357...
You're right
sorry about that
manic glacierBOT
edgy flax
@idle owl @lone axle looking to get involved in the community as a give-back. See a bunch of first issues that are just annotation fixes. Git i've worked though, but stalled at the walk-through at https://github.com/adafruit/Adafruit_CircuitPython_AdafruitIO/issues/79 (sorry). Any guidance or ability to get me started?
lone axle
<@330227457296957440> <@382939733107408897> looking to get involved in the comm...
I published a video that walks through the process of adding the typing annotations. https://www.youtube.com/watch?v=ByOiVKnJIMI if you haven't seen that yet it's a good spot to start to get a feel for how to look through and figure out the types as well as add them into the code.
edgy flax
thanks!
lone axle
AdafruitIO is one of the larger libraries. If you're interested there are probably some smaller ones that you could attempt first. Although if you have experience with the library that can help in figuring out the types so if you're already familiar with AdafruitIO library and want to jump in on it that is cool as well.
idle owl
AdafruitIO is one of the larger libraries. If you're interested there are probab...
@edgy flax I agree with this entirely. I hosted development sprints at PyCon 2022, and everyone who picked up a type annotation issue managed to pick a super difficult one. They persevered, but it was a lot of work for a first submission. I would consider starting with a simpler one to get your feet wet, and once you're comfortable with the process, return to a more difficult library. 🙂
edgy flax
also thanks for that heads up. I noticed that the line numbers given in the incidents also don't really line up anymore (but only one or two off, so no biggie). Perhaps i'll look for a smaller one to get going on.....
idle owl
If something changed in the library, then yes, things won't match up. We made those issues last year sometime, so things have almost certainly changed in many of the libraries.
proven garnet
Is it worth removing the label from some of the more difficult ones? Looking at you, adafruit_ble.
idle owl
Is it worth removing the label from some of the more difficult ones? Looking at...
Absolutely. If you'd like to go through them, that would be amazing.
proven garnet
The ones that come to mind are Adafruit_BLE, Adafruit_IO, Adafruit_GPS (I recall it being wonky). Should I just do all the BLE libraries?
unique python
I recently contributed with a couple of type annotation issues, and it was fun. But yeah, there are some difficult ones. It makes sense to first scan the code a bit to see how it looks like, before jumping in 🙂
idle owl
The ones that come to mind are `Adafruit_BLE`, `Adafruit_IO`, `Adafruit_GPS` (I ...
Not all of them are as bad as the main one, I don't think.... but who knows. Make a judgement call 🙂
idle owl
@lone axle Does this code only check main, or what's the deal there? I pushed a branch, and it doesn't seem to be picking up the branch run.
To my fork, and I updated the URL to be my account.
This is what I came up with for testing purposes. Push/PR within my own fork.
Well, what Jeff suggested anyway.
lone axle
The code bases what it checks on the URL variable near the top.
lone axle
I did add some comments last night and pushed those to my branch as well. https://github.com/FoamyGuy/Github_Actions_Status_Light/blob/main/github_actions_status_watcher.py
idle owl
Thank you!
lone axle
Hmm, I'm not sure about the branches honestly. I did all of my testing on the main circuitpython repo.
idle owl
Ok
lone axle
one moment. We can pull the same data in the browser and get an idea of what the script will see
idle owl
I looked at that JSON file from the API link, and it doesn't show my branch push.
idle owl
Current ID is 2584666641
I didn't see it in the file.
Or is it better to only show main? That would handle PRs which is I guess what matters.....
But maybe not for everyone?
Bleh.
lone axle
Oh, well one thing I think is it's looking specifically for the build.yml workflow. It doesn't seem like your push triggered that workflow to run
idle owl
Right... where even is that? I may have deleted it when purging workflows for testing. Which would explain why it didn't run.
NM, I did.
I'll add that back and try again.
lone axle
the specific workflow is specified toward the end of the URL as well. If you wanted to use it for pre-commit.yml instead of build should be able to use this URL:
https://api.github.com/repos/kattni/circuitpython/actions/workflows/pre-commit.yml/runs
Oh fair enough.
Maybe that's enough for testing. Though I wanted to show the GitHub end of things failing next to the light telling you it failed, so..... hmm.
I'm already creating an artificial environment here. Not sure how artificial I want to take it.
Hmm. Changed it to pre-commit.yml and it shows this in serial console: Fetching workflow run status. completed - failure adding 2584666641 to shown IDs
But the light does nothing.
Ok, now it's finding build.yml (I added it back)
The light is misbehaving though. We can discuss it when we meet up in 25 minutes.
idle owl
@lone axle Ok, ready when you are. Good to vid?
lone axle
<@382939733107408897> Ok, ready when you are. Good to vid?
same, yep.
edgy flax
windows. Pre-commit run in a command line or git-bash? Pip says it's installed but it's not running from a command line......'pre-commit' is not recognized as an internal or external command,
operable program or batch file.
devout jolt
Hi! I just did a fresh checkout of the circuitpython repo, set everything up and did my standard QTPyM0 build to test and got this:
% cd ports/atmel-samd && make BOARD=qtpy_m0
Use make V=1, make V=2 or set BUILD_VERBOSE similarly in your environment to increase build verbosity.
during GIMPLE pass: evrp
../../py/gc.c: In function 'gc_mark_subtree':
../../py/gc.c:1178:1: internal compiler error: Illegal instruction: 4
1178 | }
| ^
Please submit a full bug report,
with preprocessed source if appropriate.
See <https://bugs.linaro.org/> for instructions.
make: *** [../../py/mkrules.mk:62: build-qtpy_m0/py/gc.o] Error 1
Am I forgetting something?
devout jolt
<@352910176736772096> did you try a clean?
yeah this is a fresh checkout in new directory.
% arm-none-eabi-gcc --version
arm-none-eabi-gcc (GNU Toolchain for the Arm Architecture 11.2-2022.02 (arm-11.14)) 11.2.1 20220111
looking on how to backrev in homebrew
slender iron
I think the CI is on GCC 10 still
I'm on 12 though and I think its ok
are you on an arm mac?
devout jolt
how 12? isn't the current 11? https://developer.arm.com/tools-and-software/open-source-software/developer-tools/gnu-toolchain/downloads
slender iron
I use the package from Arch Linux
devout jolt
yes Arm Mac but never had a problem before on this box
idle owl
Checking
devout jolt
yeah weird, arm-none-eabi-gcc 9.2.1 works
(that's the one packaged with Adafruit Arduino)
Thanks ARM for vending out a bad compiler. I don't even see a way of getting at alternate versions from their site
slender iron
they relatively recently switched where to download it from
idle owl
I'm using a brew cask.
devout jolt
me too
slender iron
you should be able to do arm-none-eabi-gcc --version
idle owl
I did that. Oh I typoed it. 🙄
slender iron
(I may have typoed it too)
idle owl
@devout jolt arm-none-eabi-gcc (GNU Arm Embedded Toolchain 10-2020-q4-major) 10.2.1 20201103 (release)
devout jolt
the cask I'm using is referencing 11.2 (https://github.com/Homebrew/homebrew-cask/blob/HEAD/Casks/gcc-arm-embedded.rb) which cask are you using?
idle owl
gcc-arm-embedded. Maybe I pinned it somehow? Or have managed to avoid Homebrew updating it?
devout jolt
interesting. okay thanks. problem solved for now, in that it's not my problem nor CirPy's problem. 🙂 For now I'll just lamprey on my Arduino install for its Arm compiler
idle owl
Sounds good 🙂
edgy flax
updated the path and at least pylint and black run...
idle owl
@lone axle Remembered one of the things I wanted to ask. Does the .env.sh file need to be in the same directory as the .py file?
lone axle
<@382939733107408897> Remembered one of the things I wanted to ask. Does the `.e...
I don't think it has to be in the same directory, but you do have to source it to get it activated so that the script will be able to access its variable
lone axle
so if it were in a different directory you'd have to cd after activating to wherever the script is
idle owl
Ohhhh
idle owl
That makes sense. So it's a convenience thing.
lone axle
yep
idle owl
Got it. Thank you!
idle owl
@lone axle Dumb question, mostly verifying something I'm second guessing now... Numbers beginning with 0x are hex right?
lone axle
<@382939733107408897> Dumb question, mostly verifying something I'm second guess...
correct.
idle owl
Thanks!
With f-strings.... I know they take variables, but can you also add text inside the "" around the variables, and have it print the text with the parsed variables as well?
There are only variables in this f-string, so I'm not sure.
@lone axle For this comment: ```py
if program is running from CLI
if name == 'main':```
Where else would the code be running from?
jaunty juniper
an import
it's way to run code when you call python the_script.py but not when you import it
jaunty juniper
typically you write a library and have that code be run to test it by calling the main functions with the default parameters
idle owl
Hmm
jaunty juniper
but it does not interfere with using the library from another file
idle owl
Fair enough
jaunty juniper
or the other way round, you write a command line tool where everything is in functions, and the __name__ clause starts the whole thing, while reserving the possibility to call each part independently from another script, for unit tests for example
proven garnet
If I understand your question correctly:
my_name = "Tekktrik"
sentence = f"Hello! My name is {my_name}!"
print(sentence)
# Prints:
# Hello! My name is Tekktrik!
@idle owl 
idle owl
If I understand your question correctly:
```python
my_name = "Tekktrik"
sentenc...
Yes, that's it! Thank you!
devout jolt
<@382939733107408897> For this comment: ```py
# if program is running from CLI
i...
If you have any code in a .py (not in a class), it will get run on import. In normal Python you prevent this with that if-clause
idle owl
If you have any code in a .py (not in a class), it will get run on `import`. In ...
But only if you're importing the specific file right?
I guess, further, what difference does it make? Why do you want to prevent it from happening?
devout jolt
I guess, further, what difference does it make? Why do you want to prevent it fr...
Yeah. I've seen Python libraries that are also command-line tools. If you do "./foo.py", it runs and is a program, but you can also "import foo" and use foo's classes to do stuff
jaunty juniper
mostly making a library that can be quickly run as a command line tool, or for testing purposes
idle owl
So...... is it not needed here?
jaunty juniper
where is here ?
idle owl
It's this but with a bunch of tweaks I made: https://github.com/FoamyGuy/Github_Actions_Status_Light/blob/main/github_actions_status_watcher.py
idle owl
This is desktop Python
It runs directly from my computer.
I'm concerned about trying to explain it in the guide. Because I'm struggling to understand that line.
Foamyguy is going to give this page a look after I add what I can, so in theory he can fix my terrible explanation of it, but still.
devout jolt
Yes, I think it's not needed. But may be considered "good Python"? I dunno 🙂 but if you wrapped up all your code into a class and the only part of the file that actually got run was something like status_watcher = StatusWatcher(); status_watcher.run(). Then you could put that line being the if-main clause
idle owl
Hmm alright
slender iron
generally a library only defines classes and functions instead of "running code"
the __name__ == "__main__" bit allows you to store some code to run in what would normally be a library
devout jolt
agreed. but I've seen a lot of Perl/Python/Java code that embeds the __main__ stuff to let you play without needing to create a second file
idle owl
I was heading in the right direction for a bit there, but I'm getting lost again now.
devout jolt
As an example, here's the http.server module in CPython. You can use it with import http.server. You can also say python -m http.server 9000 https://github.com/python/cpython/blob/3.10/Lib/http/server.py#L1261
https://docs.python.org/3/library/http.server.html
idle owl
So in this case, this code, as-is, would not work with import but only python ...
devout jolt
the code you linked? Yes, it's just a program, not a library-that-can-also-be-a-program. (surely there's a name for that)
idle owl
Wait, so that means I should leave it in there? Because it's not the second thing?
Or do I not care in this case?
That's what I'm trying to understand at this point, I think.
devout jolt
in this case you do not care, because (I think) no one is going to be doing import github_actions_status_watcher
idle owl
Ok, yeah, that's what I figured too.
tulip sleet
@slender iron is there a .env flag that says "don't connect on startup", so that we can use .env instead of secrets.py, but not have wifi start up immediately? I wonder if maybe that should be the default...
tulip sleet
I'm just thinking that for many circumstances web workflow is not needed, but .env is very convenient
proven garnet
@idle owl the scripts I wrote for adabot to check the CI status (before I decided to make them purely command line operable) were designed to be imported into other scripts (so parts of it could be used to build more complex tools) as well as standalone (you could just run the run the python file via python ci_status.py and it would check the CI actions. But I would want to be doing a check on all the libraries for the first case, so I hid the actual "doing" when being run via python ci_status.py behind if __name__ == "__main__": so it ONLY is run when it's NOT being imported. That way imports grab the functions, classes, etc. and move on, whereas the other way also grabs them and then runs that code where it actually does a check.
idle owl
Hmm.
idle owl
<@330227457296957440> the scripts I wrote for adabot to check the CI status (bef...
I sort of follow? 😄
It's alright. I know what I need to know for this guide. Which is good enough for today. 🙂
Food time! BBIAB.
manic glacierBOT
idle owl
@lone axle I won't be able to finish this walkthrough in time for tonight. I verified that you looking at it tomorrow evening worked for Limor, she said it's fine. So I'll get it finished up tomorrow and you can take a look at it then.
lone axle
Sounds good
thorny jay
Pico W
warm stump
none of the deep sleep functions I've ever used actually work, always had to wait for event manually and only transmit when the timer runs out.
for when it does, you need alarma. 🦙 or sleepio, either one
Macropad: 0.095A with cyan lights on, 0.040A with no lights on, 0.040A in a crashed VM state under deep sleep (in safe mode)
slender iron
@warm stump the best deep sleep support is on esp
Rp2040 isn’t really designed for low power
onyx hinge
@slender iron I'm looking at https://github.com/adafruit/circuitpython/pull/6486 and https://github.com/adafruit/circuitpython/issues/6484 in relation to https://github.com/adafruit/circuitpython/issues/6523 -- trying to understand what I should do about an rgbmatrix display before entering a true deep sleep, vs what should happen for any other kind of display.
Is there any kind of display for which doing a "release displays" prior to truly entering deep sleep is a mistaken idea?
warm stump
<@539881446756843560> the best deep sleep support is on esp
Adafruit Feather from before time began is also pretty good, 0.006A whereas ESP32 was 0.025A nominal
^^ the one that now runs CircuitPython7 with a CIRCUITPY Drive on USB
slender iron
<@252717193496756235> I'm looking at https://github.com/adafruit/circuitpython/p...
eink should preserve the image
I wouldn't expect a matrix to continue during deep sleep
slender iron
thanks @tender marsh! ☝️
tender marsh
thanks <@137254691242377216>! ☝️
i did it! thanks 😄
slender iron
@tulip sleet I'm mixed about that. user code could have a different variable for SSID and password
everything with the CIRCUITPY_ prefix should be for internal use
tulip sleet
<@329766224093249548> I'm mixed about that. user code could have a different var...
well, there is likely to be only one network of interest visible? I agree, there could be another pair of SSID/password for user code.
But it would still be nice to enable/disable web workflow without having to delete the values or comment them out.
CIRCUITPY_WEB_API_SSID vs plain SSID ?
slender iron
whatever they want to call it
we should only use CIRCUITPY_ for internal stuff though
it'd be interesting to sort all of the library repos by stars
and to track the sum
edgy flax
I published a video that walks through the process of adding the typing annotati...
one type I did not see in the video that seems pretty common, pin. Is there an object type peculiar to that (microcontroller.Pin, perhaps?) or is it just 'int'? As in this
def init(
self,
pin,
series_resistor,
lone axle
one type I did not see in the video that seems pretty common, pin. Is there an...
It may depend on the specific way it's used but I think generally microcontroller.Pin is likely to be the correct type in most places.
lone axle
If you find a usage of the function and it passes in something like board.D2 or similar those are microcontroller.Pin type objects I believe.
manic glacierBOT
idle owl
@lone axle Ok, so here's the situation. I made some changes to the code, and I need to get it into the Learn repo so I can get it embedded in the guide. This is obviously happening before you have a chance to look at my changes and the walkthrough. So! Here's my thought. I'll submit it as-is, the walkthrough applies to my version of the code. Then, if you want to make code changes, you'll put in a PR to the Learn repo, and update the walkthrough page to match when you get to it this evening. And if Anne doesn't beat me to it, I'll get it merged tomorrow so the guide is matchy. Does that work?
lone axle
okay, yep. Sounds good
idle owl
Thank you!
idle owl
Oof. @lone axle The URL line is getting flagged too long by Pylint. Should I split the string, make it less readable, or disable Pylint for that line? Not sure either of those options is great.
lone axle
Personally I would lean toward a pylint ignore for the length of it. That way the URL can be copy/pasted from the browser which is very helpful for verifying the raw data that the script is seeing.
lone axle
If it does get split up it could be like github_username as one var, repo_name as one var, and workflow_name as one var. Then concatenate them all to make the full URL. Which would still be clear code I think, but less convenient to copy/paste.
idle owl
Using with ti open a file in Python, this documentation seems to indicate it closes automatically? Does that mean I don't need to call f.close()?
idle owl
If it does get split up it could be like github_username as one var, repo_name a...
I like the idea of an intact URL.
lone axle
That is correct as far as I understand. When you use with to open the file. It'll get automatically closed whenever the block inside the with ends.
warm stump
ESP32-S original with the factory firmware update tool, reliable serial from the 16u2 on an Uno (would probably also work with a Metro or Mini) without having to remove the AVR
this is where SoftwareSerial will fail and the discontinued PL2303 products also have failed. Tried uploading the ESP32-S2 CircuitPython to it, it froze while reading the stub for a while
probably knowing it's an incompatible board, but it looks like there is still some life in it. Would require the USB on e.g. SAMD21 or 51 to bring it up to par with the latest boards
the one in the image is DM, the one I have is Espressif
manic glacierBOT
fluid needle
is there a circuitpython uf2 for the Pico W?
stuck elbow
no
fluid needle
ahh would have been too good ^^ I guess we'll have to wait a bit
stuck elbow
possibly quite a while, especially if you want the wifi support too
fluid needle
oh yeh, would love to do some HID stuff over wifi with adafruit_hid lib
fluid needle
ah no i mean HID via USB but control it via wifi
slender iron
the ESP32-S2 or S3 can do that already
slender iron
ESP32-S original with the factory firmware update tool, reliable serial from the...
I think the chip on this is an ESP32 (no S)
proven garnet
@idle owl I wrote the script for parsing all the requirements for the Bundle libraries and ran it. I've got a list of requirements for each library ready to go for the next step of creating pyproject.toml files! Also resolved any disagreements between setup.py and requirements.txt while doing so.
It's worth talking about whether we want to use optional dependencies for some libraries with extra functionality not immediately enabled, like using pillow for specific image related functionality, or for when example files use something but the actually library does not (one example I saw of this was pyserial for an example). There's merit to splitting things up into what's needed at the minimum, but I also see hiccups when people are confused why things like examples won't run.
In that case it may just be worth defining everything as a regular dependency
deft jay
Question: Are all CircuitPython boards supposed to show up as USB Flash Drives? I'm reading up on the ESP32-C3 board, excited about RISCV, very cool. Not sure what it means that it doesn't show up as a disk drive though. I thought all Circuitpython ports/boards were supposed to show up as drives as a term of their inclusion into the project. How did ESP32-C3 be made an exception for this rule? I don't have one on hand to play with, so I haven't confirmed that it doesn't show up as a disk drive. Just curious about the project development decisions. Not so much a technical question as much as a project question.
slender iron
Question: Are all CircuitPython boards supposed to show up as USB Flash Drives? ...
As of CP 7 we also allow a BLE workflow. We thought C3 would have that but ran into a snag. In 8 we're introducing a (just merged) web workflow too.
Just added this page to the docs: https://docs.circuitpython.org/en/latest/docs/workflows.html
warm stump
I think the chip on this is an ESP32 (no S)
it seems to revert to 2016 so that seems about right. boards like this with USB to UART adapters might not show up as USB Drives, it ideally needs a dedicated USB device peripheral like the RP2040 has. would be ideal for a pico with no Infineon WiFi on-board
slender iron
Right. Dan is working on ESP32 support now actually. Since we're adding the web workflow
deft jay
very very cool!
slender iron
yuuuup. I'm excited about it
deft jay
I'll have to read up more on the web workflows. Is that a way to update scripts without touching the boards? That could be very useful.
slender iron
yup! it is a REST API
you can just do a PUT to circuitpython.local/fs/code.py to update
deft jay
Link for the lazy on where to read up on that? Also is pico W support anticipated soon?
slender iron
this link has the API docs: https://docs.circuitpython.org/en/latest/docs/workflows.html
soon no. eventually yes
gotta get it all going on ESP first
(and I'm out for 12 more weeks of paternity leave in 6 weeks or so)
deft jay
jealous. I just had my first. I went back after a week. Zero paternity leave at my company. I have no vacation left. Enjoy it and congrats!
slender iron
oh no! thats a bummer. how old are they now? mine is just over 3 months
(the last 6 of mine are unpaid)
fluid needle
gz to the Human Pico 😄
deft jay
She's 10 weeks old. I've worked for 8 of them 😦
So, I'm curious, is esp32-c3 web workflow functional right now?
jaunty juniper
I don't think I got it to work on a QT PY C3 with a previous commit, but with the REPL issues on the QT PY it's already hard to even create the .env file...
onyx hinge
does wifi workflow do repl over wifi yet (once configured)? or is repl still over serial?
jaunty juniper
should we have a quick 7.3.2 for the matrixportal traceback module ?
tulip sleet
is there a circuitpython uf2 for the Pico W?
The regular Pi Pico build should work, but no wifi support.
tulip sleet
should we have a quick 7.3.2 for the matrixportal traceback module ?
I was thinking about that, seeing if anything else showed up.
jaunty juniper
except board.LED is not valid on the pico W
(the LED goes through the wifi module)
tulip sleet
does wifi workflow do repl over wifi yet (once configured)? or is repl still ove...
Scott is working on websocket support for serial
onyx hinge
nice!
tulip sleet
@slender iron would an ESP32SPI-style CircuitPython library be possible with the Pico W?
tidal kiln
@tulip sleet do you know if there are any HTTPServer guides in work? (for the CP lib)
tulip sleet
<@329766224093249548> do you know if there are any HTTPServer guides in work? (f...
no, partly because I'm not satisfied with the library
tidal kiln
oh. it worked for me 🙂
was curious about it, so tried this last nite
import wifi
import socketpool
from adafruit_httpserver import HTTPServer, HTTPResponse
print("Starting access point.")
wifi.radio.start_ap("QTPY_ESP32S2")
print(f"Listening on http://{wifi.radio.ipv4_address_ap}:80")
pool = socketpool.SocketPool(wifi.radio)
server = HTTPServer(pool)
@server.route("/")
def base(request):
print("GET base")
return HTTPResponse(filename="/index.html")
@server.route("/button1")
def button1(request):
print("Button 1")
return HTTPResponse()
@server.route("/button2")
def button2(request):
print("Button 2")
return HTTPResponse()
@server.route("/button3")
def button3(request):
print("Button 3")
return HTTPResponse()
@server.route("/button4")
def button4(request):
print("Button 4")
return HTTPResponse()
server.serve_forever(str(wifi.radio.ipv4_address_ap))
serve up some basic buttons and can connect and control via a smartphone
tulip sleet
see the issues for some quirks. It definitely works! I was just thinking there could be more features to add, possibly with API changes. If you are interested in doing a tutorial or project guide, I won't object 🙂
tidal kiln
it can wait. was just something i was toying with.
but seems like it could make for cool guide at some point.
very generic framework that could be used for anything
<html>
<head>
<script type="text/javascript">
function postIt(route) {
var xmlreq = new XMLHttpRequest();
xmlreq.open("GET", "http://" + location.host + "/" + route)
xmlreq.send();
}
</script>
</head>
<body>
<center>
<button onclick=postIt("button1")>BUTTON1</button><br/>
<button onclick=postIt("button2")>BUTTON2</button><br/>
<button onclick=postIt("button3")>BUTTON3</button><br/>
<button onclick=postIt("button4")>BUTTON4</button><br/>
</center>
</body>
</html>
there's the basic html that went with it
warm stump
https://github.com/TheMindVirus/httpsserver/tree/websockets - same, I don't like websockets either
plus a little stability for deep sleep on ESP32-S for Pico H, not sure why not doing this would lead to instability on Serial
tulip sleet
```html
<html>
<head>
<script type="text/javascript">
function postIt(route) {
...
it was originally written for a CPy version of https://learn.adafruit.com/wordle-personal-esp32-s2-web-server/
i like the AJAX thing you are doing
the routing could be better, no explicit arg passing right now
tidal kiln
still amazing how simple it was to get it working
could control LEDs, motors, whatever
wraith crow
I've built the latest bits for an UM FeatherS2 and have been trying to connect via the web workflow, so far I'm just getting timeout errors. I am able to ping the IP address. I've got the .env file with CIRCUITPY_WIFI_SSID, CIRCUITPY_WIFI_PASSWORD AND CIRCUITPY_WEB_API_PASSWORD installed and the board does get an IP address on it's own when it powers up. At one point during testing I did get the API password prompt when attempting to reach the /fs/ page but after entering the password it never continued and I haven't been able to get that far again.
Any chance I'm missing something basic? Is there a way I can get any debugging info out of the workflow?
tidal kiln
@tulip sleet can come back to it later when ur happy with the library. was just a quick fun demo thing for now 🙂
deft jay
the guide seems to be written for esp32-S
slender iron
I've built the latest bits for an UM FeatherS2 and have been trying to connect v...
debug builds will output to the debug uart. does the root page work? it isn't password authed
slender iron
<@252717193496756235> would an ESP32SPI-style CircuitPython library be possible ...
Possible yes. Its hard to write any library for it since details are behind an NDA
tulip sleet
the protocol is NDA?! oy
slender iron
So, I'm curious, is esp32-c3 web workflow functional right now?
its enabled but I've been testing on S3. I haven't checked C3
slender iron
the protocol is NDA?! oy
yuuuuup
wraith crow
debug builds will output to the debug uart. does the root page work? it isn't pa...
@slender iron No I haven't gotten the root page to come up yet, does having the usb serial connection active interfere with it? I'll go ahead and try a debug build this evening 🙂
tulip sleet
i have reworked the REPL connection to be a "console" UART, rather than "debug" only
as part of the ESP32 port
wraith crow
warm stump
Possible yes. Its hard to write any library for it since details are behind an N...
I'd find it funny if the reason that it was behind an NDA was that it was secretly just translating all your calls to AJAX in the backend instead
wraith crow
I'm thinking I didn't pay enough attention to this line:
The web server will allow requests from cpy-XXXXXX.local, 127.0.0.1, the device's IP and code.circuitpython.org. (circuitpython.local requests will be redirected to cpy-XXXXXX.local.)
Does that mean I need to give a cpy-XXXXXX.local alias to the machine I'm trying to reach the microcontroller from?
tulip sleet
cpy-XXXXXX.local is the board itself. you can make requests from any host computer. Maybe that should be rephrased
^ @slender iron
slender iron
I did have a code.py, but just to be sure I moved it out of the way. Do I need ...
no, you shouldn't. but there might be some that interfere
slender iron
cpy-XXXXXX.local is the board itself. you can make requests from any host comput...
ya, its weird because its AJAX requests from particular pages. (The Origin header)
wraith crow
Ok, I've updated my host file and tried the example curl connect:
C:\Users\Desktop\curl-7.84.0_3-win64-mingw\bin>curl -v -X OPTIONS -L --location-trusted http://circuitpython.local/
- Trying 10.0.0.187:80...
- Connected to circuitpython.local (10.0.0.187) port 80 (#0)
OPTIONS / HTTP/1.1
Host: circuitpython.local
User-Agent: curl/7.84.0
Accept: /
And then it just sits.....
I've got to run some errands but when I get back I'll try building on an S3 and then try a debug build.
wraith crow
Yea, I can ping the 10.0.0.187 and that's what shows on the console title bar
slender iron
try curl -v http://circuitpython.local/
wraith crow
C:\Users\Desktop\curl-7.84.0_3-win64-mingw\bin>curl -v http://circuitpython.local/
- Trying 10.0.0.187:80...
- Connected to circuitpython.local (10.0.0.187) port 80 (#0)
GET / HTTP/1.1
Host: circuitpython.local
User-Agent: curl/7.84.0
Accept: /
slender iron
hrm
wraith crow
Are you using the second core for the web workflow?
slender iron
no, it should be the same core as the rest of circuitpython
try resetting the board and doing the GET / again
wraith crow
@slender iron same result
slender iron
hrm, ya. A debug build and its output would be helpful to see
does the serial connection respond still?
wraith crow
Okay, I'll pick it up later this afternoon
Yes, the serial USB connection stays functional throughout
slender iron
so its not too fatal of an issue 🙂
manic glacierBOT
obsidian jasper
Hi, all!
Especially, Scott)
I want to ask - do you have source files with working implementation of usb host(imxrt) with new tinyusb lib? I'm asking because I saw that you have done with working usb keyboard...
idle owl
<@330227457296957440> I wrote the script for parsing all the requirements for th...
Great!
idle owl
It's worth talking about whether we want to use optional dependencies for some l...
Yeah fair enough. Monday maybe.
fluid needle
Neradoc aren't you the guy who supplies us with CP hid keyboard layouts? 
jaunty juniper
Possible yes. Its hard to write any library for it since details are behind an N...
oh yeah I flinched when I read that in the release news post:
For C developers, today’s release of the Pico SDK includes wireless networking support. The network stack is built around lwIP, and uses libcyw43 from Damien George (of MicroPython fame) to communicate with the wireless chip. By default, libcyw43 is licensed for non-commercial use, but Pico W users, and anyone else who builds their product around RP2040 and CYW43439, benefit from a free commercial-use license.
is that a library we can include in CP on the board ?
slender iron
ya, we can include it and use it for RP2040 builds
jaunty juniper
Neradoc aren't you the guy who supplies us with CP hid keyboard layouts? <:kingl...
slender iron
I know damien put a lot of work into it. so its his call
fluid needle
that <https://github.com/Neradoc/Circuitpython_Keyboard_Layouts>
oh yeh saved me a lot of time 😄 many thanks ❤️
proven garnet
Yeah fair enough. Monday maybe.
Sounds good, not sure if you want to do it on the 4th though. Happy to do it with a grilled hotdog in hand though haha 🌭
jaunty juniper
@slender iron I'm looking at adding setting up the default BLE name of a board using dotenv, what do you think ?
daring pumice
Question about building CP for a custom board: My design is heavily based on the Feather Sense, but has some notable changes to GPIO assignments (namely the primary LED and Reset pins).
So far I've tried to make changes to the makefile and pins.c in /ports/nrf/feather_bluefruit_sense - is that the right place?
jaunty juniper
yep, you have to look at mpconfigboard.h for the pins of the default busses (board.I2C, etc.) and other setup like the NEOPIXEL and LED pins
daring pumice
Cheers!
To make sure I'm changing things correctly, if I want to change the primary LED (for example) to GPIO P1.15, I would need to change these lines:
{ MP_ROM_QSTR(MP_QSTR_L), MP_ROM_PTR(&pin_P1_09) },
{ MP_ROM_QSTR(MP_QSTR_LED), MP_ROM_PTR(&pin_P1_09) },
{ MP_ROM_QSTR(MP_QSTR_RED_LED), MP_ROM_PTR(&pin_P1_09) },
to
{ MP_ROM_QSTR(MP_QSTR_L), MP_ROM_PTR(&pin_P1_15) },
{ MP_ROM_QSTR(MP_QSTR_LED), MP_ROM_PTR(&pin_P1_15) },
{ MP_ROM_QSTR(MP_QSTR_RED_LED), MP_ROM_PTR(&pin_P1_15) },
Correct?
jaunty juniper
yep, you can ditch the L name, it's just for backwards compatibility on that board, and RED_LED too if you only have one (and it's not red)
you want to make the same change to MICROPY_HW_LED_STATUS in mpconfigboard.h too
daring pumice
you want to make the same change to `MICROPY_HW_LED_STATUS` in mpconfigboard.h ...
Have that done 🙂
daring pumice
yep, you can ditch the L name, it's just for backwards compatibility on that boa...
Awesome. I'll do that 🙂
I also need to define the reset pin
slender iron
<@252717193496756235> I'm looking at adding setting up the default BLE name of a...
Fine with me!
daring pumice
I'm not seeing that in either mpconfigboard.h or pins.c
Maybe I'm looking for the wrong name?
jaunty juniper
the reset pin is usually hardware, it's not a thing you define in software, is it ?
daring pumice
I need to soft reset (I think) so I can signal a "double click" of the reset pin.
Not sure if I'm being clear...
sry
tulip sleet
just to be clear, the reset button is the real hw reset line. THere's no soft reset pin
daring pumice
Awesome.
I won't worry about that 🙂
Gonna build this and report back 😉
Thanks so much @jaunty juniper and @tulip sleet
idle owl
Sounds good, not sure if you want to do it on the 4th though. Happy to do it wi...
Tuesday! 🙄
daring pumice
Ok. I've set the pins inside the appropriate header files, but I'm not seeing the primary LED light.
Could I be missing something?
slender iron
it is MICROPY_HW_LED_STATUS in mpconfigboard
daring pumice
That's pointing at the same pin
P1.15
#define MICROPY_HW_LED_STATUS (&pin_P1_15)
wraith crow
Web workflow testing
lone axle
how does the welcome_html variable used here: https://github.com/adafruit/circuitpython/blob/8814ee03f1c716e7b7ee812799c83893b9e5d436/supervisor/shared/web_workflow/web_workflow.c#L990 come to exist so that it can be passed to _REPLY_STATIC() ? I found this section earlier in the code that seems related: https://github.com/adafruit/circuitpython/blob/8814ee03f1c716e7b7ee812799c83893b9e5d436/supervisor/shared/web_workflow/web_workflow.c#L827
I tried creating a new one of those though edit_html and I'm unable to get a build to complete, having errors about undefined reference to 'edit_html'
I haven't been able to find any other set up that is occuring for welcome_html and the other static files. But I feel like I must be missing some from somewhere.
slender iron
lone axle
I did create a line similar to those:
STATIC_FILE(directory_html);
STATIC_FILE(directory_js);
STATIC_FILE(welcome_html);
STATIC_FILE(edit_html); // <- added this one
STATIC_FILE(welcome_js);
STATIC_FILE(blinka_16x16_ico);
what is the size limit of the BLE device name ? I believe the total limit is 31 bytes for the scan response, but it contains additional stuff in it, and according to some googling the practical limit would be around 21 or something ? Or is it automatically cut anyway by the underlying BLE stack so I don't need to bother too much ?
manic glacierBOT
slender iron
what is the size limit of the BLE device name ? I believe the total limit is 31 ...
it depends on the other stuff in the scan response
tiny token
see the issues for some quirks. It definitely works! I was just thinking there c...
I like the idea of adding features to the HTTPServer lib and making it more robust. I don't have a lot of time to assist but could provide some help.
I was trying to look at others or even roll my own, but haven't gotten very far.
tulip sleet
I like the idea of adding features to the HTTPServer lib and making it more robu...
In the issues for the library I mention some other libraries that might be starting points.
tiny token
Any idea why there is the HTTPServer and also the WSGI server? seem to be trying to implement similar things.
tulip sleet
I think someone mentioned ampule too. Not sure if that is in the issues.
jaunty juniper
ampule looks quite nice
tiny token
I was looking at ampule last night. Seems very similar to HTTPServer provides some better routing with args
tulip sleet
The WSGI server didn’t serve files natively. That was my primary goal for HTPPServer
jaunty juniper
I'd like to look at the differences with the wsgi library and merge them into the HTTPServer library, if there's anything worth it but I'm not gonna be able to make time for that
tulip sleet
HTTPServer is very new and has no legacy of use. I’m happy to completely change it or replace it with something else
and then archive it
I am not familiar with the libraries in this space, so if we can base a new library on something more known, that's fine
tiny token
do you know if the WSGI standard provides for native files? it'd be nice to create something that is standards based if possible
jaunty juniper
it needs a polling version that allows running other things in the loop, or from asyncio
tiny token
tiny token
it needs a polling version that allows running other things in the loop, or from...
I agree and that's an easy fix. I've played with that
tulip sleet
I think the assumption might be that wsgi is fronted by nginx or some other server that serves the static files and passes on dynamic requests to wsgi
that is the impression I get from websearching on the topic
jaunty juniper
and I don't know what should be done with https://github.com/adafruit/Adafruit_CircuitPython_WSGI
tulip sleet
I have a partial draft of a version of adafruit_requests with some async API, but I haven't had time to work on it further recently
your contributions to this would be most welcome
tiny token
splitting out parts of the current code in serve_forever would provide for a start and a poll function.
tulip sleet
earlier today @tidal kiln (who works for Adafruit) was talking about doing AJAX calls with the current HTTPServer; maybe you saw that already: about 7.5 hours ago
I said pretty much the same thing then -- no problem changing the current API
jaunty juniper
it depends on the other stuff in the scan response
interestingly, if I pass a string that's too long, it defaults to "nRF5x"
slender iron
have a great long weekend everyone!
(non-US folks can take monday off because you all are awesome)
😛
tiny token
your contributions to this would be most welcome
created a pull request for the split of serve_forever https://github.com/adafruit/Adafruit_CircuitPython_HTTPServer/pull/7
integrity should still be there for the existing method so it doesn't break any existing code using it.
eventually it'd be nice to merge the WSGI code and this code into a more robust web server, but that's something for another day. Cheers!
manic glacierBOT
tiny token
My pull request keeps failing CI with this "reformatted" error, but I don't understand what it's telling me. Any words of wisdom?
blissful pollen
My pull request keeps failing CI with this "reformatted" error, but I don't unde...
It is failing the black code formatting check. It is kinda confusing cause the error says reformatted but it does not actually reformat it.
Do you have pre-commit hooks installed on your PC?
tiny token
I just found that article. am installing it now. (https://learn.adafruit.com/creating-and-sharing-a-circuitpython-library/check-your-code)
blissful pollen
That should fix it
tiny token
that'll fix it just by running it, or it'll point me to the issue?
blissful pollen
once you run the pre-commit hooks (you probably have to run it manually since you already committed) black should reformat your files and you can commit them
blissful pollen
No problem, it still catches me sometimes
tiny token
There it is the lovely ✅
jaunty juniper
I'm getting "thing is deprecated" errors from uncrustify when I build CP
ok fixed it
rough folio
I think this is more suited for here- but blinka installed in a venv and then run through a jupyter kernel fails to start because board can't identify the platform
(on windows)
jaunty juniper
what platform should it identify ?
rough folio
I have no idea but it precludes the notebook from starting entirely which is pretty undesirable
because ipython will try to load modules
jaunty juniper
it tries to load modules that are not in use ?
rough folio
same dir
I think anyway
let me take another look at this error
yeah it seems to try and load modules in the same directory
I can put it in a subfolder I suppose
see if that does anything
yep it does I guess that isn't really an issue
I was kind of under the impression it was trying to index the venv or something and not just the modules in the same folder
manic glacierBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
ornate breach
Is the P19A variant of the SAMD51 supported in circuitpython?
I imagine so since it’s in the uf2-samdx1 library
stuck elbow
Is the P19A variant of the SAMD51 supported in circuitpython?
there are a couple of boards that seem to be using it
aloriumtech_evo_m51 and seeeduino_wio_terminal
ornate breach
Oh perfect
manic glacierBOT
digital shoreBOT
adafruit
slender iron
<@&356864093652516868> Just a reminder that the meeting is in just over 24 hours. It is a holiday here in the us today. Talk with you tomorrow!
jaunty juniper
weird error in circup's CI during pip install:
https://github.com/adafruit/circup/runs/7188851151?check_suite_focus=true
Collecting circuitpython-build-tools
Downloading circuitpython-build-tools-1.10.1.tar.gz (11 kB)
Preparing metadata (setup.py): started
Preparing metadata (setup.py): finished with status 'error'
error: subprocess-exited-with-error
× python setup.py egg_info did not run successfully.
│ exit code: 1
╰─> [25 lines of output]
Weird, but it looks like it's fixed if importlib_metadata is pinned to a new release, I tested the lasted (4.12.0) and it seems to have passed CI.
jaunty juniper
ah some part of the setup tools that doesn't manage its dependencies well maybe
how would I fix the circup PR then ?
proven garnet
I submitted a PR to update it; it looks like just re-pinning that requirement should be good.
It doesn't look like change from an API perspective for importlib_metadata.version() (the only thing we use from it, it seems) from v0.20 to v4.12!
jaunty juniper
oh wow that is a jump !
(you changed the - with a _, pip doesn't care, but the - is usually used I think)
proven garnet
Oh, good catch, I'll update that - if nothing else it's my own pet peeve as well haha
Yeah I guess something changed for setuptools since the last time the Build CI ran - It cleared successfully last time when it used setuptools==47.1.0 but not this time when it was using setuptools==63.1.0. Also a wild jump!
main furnace
Is async support for UART on the roadmap? I want to use the NeoKey1x4 and UART together.
stuck elbow
what do you need exactly?
main furnace
I've set up latching on the NeoKey1x4 keys. Right now I'm doing:
while True:
latch_update() #update the state of the NeoKey1x4
buffer = uart.read(12)
With this code you have to wait for the uart.read to return or time out before the keyboard update runs. Rather than just tapping the key to toggle it you have press and hold it.
Maybe I should check the uart for bytes available before I commit to reading it...
blissful pollen
Is async support for UART on the roadmap? I want to use the NeoKey1x4 and UART t...
It isn't anything official (or even well tested) but I did a proof of concept for UART and asyncio
https://gist.github.com/gamblor21/e9681057222f6a70fca0608ae3054b3e
It was reading GPS (via UART) there is a uart_asyncio.py in there and if you look at the GPS .py gives you an example of some of the usage.
proven garnet
oh and thanks <@135415127879974912> 👍 😉
No problem!
main furnace
It isn't anything official (or even well tested) but I did a proof of concept fo...
Thanks, I think I could do something like that. I guess a lot of libraries would need to change in order to support asyncio.
stuck elbow
it's really tacked on, both in cpython and other pythons
proven garnet
@idle owl let me know if today is still a good day to meet, I can do the usual noon or post-5pm, no preference!
blissful pollen
Thanks, I think I could do something like that. I guess a lot of libraries would...
That’s what I found and Dan had mentioned similar to me. It’s a bit more then just having to subclass the initial work you need to add the async keyword in a lot of places.
On the plus side it did work for GPS
tulip sleet
I've set up latching on the NeoKey1x4 keys. Right now I'm doing:
```python
while...
You can also set the uart timeout to zero.
thorny jay
Sorry no notes from me for today's meeting. Despite the extra day, I did not find the time to work on my project... and very soon a PicoW will arrive (and also an ESP32-S3 TFT Feather 4/2).
idle owl
<@330227457296957440> let me know if today is still a good day to meet, I can do...
I misunderstood what you meant by "discuss", I was assuming In The Weeds, but you won't be there anyway. Tomorrow at noon would work. Today does not 🙂
idle owl
<@&356864093652516868> CircuitPython Weekly happens in just under 2 hours! Please add your Hug Reports and Status Updates to the notes doc. Note "not in meeting" if you're not able to attend, or "text only" if you're not able to speak during the meeting. Looking forward to chatting with you soon!
proven garnet
I misunderstood what you meant by "discuss", I was assuming In The Weeds, but yo...
Would it still be helpful to leave notes? I'll likely be text only isntead of missing
idle owl
Would it still be helpful to leave notes? I'll likely be text only isntead of m...
Nah. I'd rather you be there. So we should chat between us, and then decide if we want to add it to In the Weeds for next Monday.
proven garnet
Sweeto!
daring pumice
Having an issue getting the LED to work with a custom PCB based on the Adafruit Feather (nrf52840).
When building and pushing the bootloader (https://github.com/adafruit/Adafruit_nRF52_Bootloader/releases/tag/0.6.4) to my board, I'm able to properly adjust the primary LED in the makefile. Because of a flaw in the design (which is being fixed in a future revision) I can't actually use the bootloader to access the board, but the status update from the LED is useful.
However, when I build and push a full version of CP to the board (using the Feather-Bluefruit-Sense port, adjusting the GPIOs to reflect my layout), the primary LED isn't getting used.
Anyone else encounter an issue like this, or have any guidance here?
slender iron
@daring pumice can you break on circuitpython main to confirm it was loaded correctly?
daring pumice
<@447309542863273985> can you break on circuitpython main to confirm it was load...
break?
slender iron
yes, with a debugger
daring pumice
I don't get a REPL, but I can halt the code using JLink Commander
when I read out the registers I get No Faults.
Unfortunately, I can't see any symbols when using GDB
slender iron
what is the backtrace when you halt?
daring pumice
Good question, I'll take a look now
jaunty juniper
oh weird I get a MemoryError after code.py ends, by creating a logging.FileHandler to an SD Card
daring pumice
what is the backtrace when you halt?
Not sure how to get a trace using JLink Commander. I can create a GDB connection, but not sure how to get a trace when I halt using the gdb CLI
slender iron
Not sure how to get a trace using JLink Commander. I can create a GDB connection...
similar process for nrf: https://learn.adafruit.com/debugging-the-samd21-with-gdb
daring pumice
similar process for nrf: <https://learn.adafruit.com/debugging-the-samd21-with-g...
Yes, I've used this to get myself this far, but this requires me to define a breakpoint somewhere in the c code - I don't know enough about where to insert a breakpoint to get a useful trace
slender iron
on main 🙂
slender iron
yup
daring pumice
Same issue as above: no symbol table :/
daring pumice
oh weird I get a MemoryError *after* code.py ends, by creating a `logging.FileHa...
Sorry @jaunty juniper , not sure if this was about my issue or something else?
jaunty juniper
no no not at all
daring pumice
kk 👍
slender iron
@daring pumice use a DEBUG=1 build
daring pumice
Sounds good. Will rebuild and get back. Much appreciated!
jaunty juniper
@proven garnet adafruit_logging.getLogger() defaults to a None handler, but that is not how C python does it, it defaults to prints (another issue is streamHandler does not end lines with "\r\n", like fileHandler does) I'll look into fixing it later unless you want to ?
daring pumice
<@447309542863273985> use a DEBUG=1 build
Hmmmm.. that doesn't seem to have worked. Same issue.
nm, was loading the .hex, not the .elf
I now have symbols
Ok. I'm now getting proper breakpoints.
daring pumice
@slender iron Here's the backtrace:
reset_into_safe_mode (reason=reason@entry=MICROPY_FATAL_ERROR) at ../../supervisor/shared/safe_mode.c:126
126 }
(gdb) backtrace
#0 reset_into_safe_mode (reason=reason@entry=MICROPY_FATAL_ERROR) at ../../supervisor/shared/safe_mode.c:126
#1 0x00073f80 in __fatal_error (msg=msg@entry=0x9f0f8 "Assertion failed") at ../../main.c:994
#2 0x0002a858 in __assert_func (file=file@entry=0xa1378 "nrfx/drivers/src/nrfx_qspi.c", line=line@entry=192,
func=func@entry=0xa1430 <__FUNCTION__.5> "nrfx_qspi_cinstr_xfer", expr=expr@entry=0xa13c4 "p_config->wipwait")
at ../../main.c:1001
#3 0x0005c820 in nrfx_qspi_cinstr_xfer (p_config=p_config@entry=0x2003fcf8, p_tx_buffer=p_tx_buffer@entry=0x0,
p_rx_buffer=p_rx_buffer@entry=0x2003fd1c) at nrfx/drivers/src/nrfx_qspi.c:192
#4 0x00068690 in spi_flash_read_command (command=command@entry=159 'Y',
response=response@entry=0x2003fd1c "ÿÿÿ£4y\003 \215ì\a", length=length@entry=3) at supervisor/qspi_flash.c:115
#5 0x0006aaa2 in supervisor_flash_init () at ../../supervisor/shared/external_flash/external_flash.c:224
#6 0x0007ec8c in flash_ioctl (cmd=<optimized out>, out_value=out_value@entry=0x2003fd34)
at ../../supervisor/shared/flash.c:168
#7 0x0008d350 in supervisor_flash_obj_ioctl (self=<optimized out>, cmd_in=<optimized out>, arg_in=<optimized out>)
at ../../supervisor/shared/flash.c:191
#8 0x000484fe in fun_builtin_3_call (self_in=0x9f214 <supervisor_flash_obj_ioctl_obj>, n_args=<optimized out>,
n_kw=<optimized out>, args=0x20012258 <_internal_vfs+56>) at ../../py/objfun.c:109
#9 0x0004ed96 in mp_call_function_n_kw (fun_in=0x9f214 <supervisor_flash_obj_ioctl_obj>, n_args=3,
n_kw=n_kw@entry=0, args=0x20012258 <_internal_vfs+56>) at ../../py/runtime.c:665
#10 0x00081234 in mp_call_method_n_kw (n_args=n_args@entry=2, n_kw=n_kw@entry=0,
args=args@entry=0x20012254 <_internal_vfs+52>) at ../../py/runtime.c:680
#11 0x00083506 in mp_vfs_blockdev_ioctl (self=self@entry=0x20012224 <_internal_vfs+4>, cmd=<optimized out>,
arg=arg@entry=0) at ../../extmod/vfs_blockdev.c:118
#12 0x00041ffc in disk_ioctl (pdrv=<optimized out>, cmd=cmd@entry=5 '\005', buff=buff@entry=0x2003fd8f)
at ../../extmod/vfs_fat_diskio.c:97
#13 0x00046fb6 in find_volume (fs=fs@entry=0x20012264 <_internal_vfs+68>, mode=mode@entry=0 '\000')
at ../../lib/oofatfs/ff.c:3142
#14 0x0007db8a in f_mount (fs=fs@entry=0x20012264 <_internal_vfs+68>) at ../../lib/oofatfs/ff.c:3395
#15 0x00046b8e in filesystem_init (create_allowed=create_allowed@entry=false, force_create=force_create@entry=false)
at ../../supervisor/shared/filesystem.c:96
I've run the code 3 times and get the same issue each time
onyx hinge
if (qspi_ready_wait() == NRFX_ERROR_TIMEOUT)
{
// This timeout should never occur when WIPWAIT is not active, since in this
// case the QSPI peripheral should send the command immediately, without any
// waiting for previous write to complete.
NRFX_ASSERT(p_config->wipwait);
return NRFX_ERROR_TIMEOUT;
}
I see that too, but further down the trace it looks like this is happening when CP is trying to mount the filesystem.
onyx hinge
It's also possible that this is a red herring; asserts are not checked except during a debug build. so your normal build might be getting past this
daring pumice
It's also possible that this is a red herring; `assert`s are not checked except ...
I hear you, but wouldn't my primary LED be getting lit at that point?
If there were no other issues, I mean?
onyx hinge
I don't know where in the start up procedure that would happen
modern wing
Good afternoon everyone -- happily lurking today.
tulip sleet
Does the bluefruit sense use an external flash?
in CircuitPython it does, but not in the bootloader
daring pumice
in CircuitPython it does, but not in the bootloader
That could be why this is failing
Its looking for a device that doesn't exist
?
My board doesn't have external flash
tulip sleet
what are you using for CIRCUITPY?
daring pumice
Good question
main furnace
:smiles: :waves:
tulip sleet
it's possible to use part of the internal flash for CIRCUITPY, but you may need the space for a larger one
proven garnet
Does it still? I thought I patched that, but maybe I didn't. Happy to review the patch when you get to it! Also good catch on StreamHandler, I didn't realize it was CRLF.
daring pumice
it's possible to use part of the internal flash for CIRCUITPY, but you may need ...
I'm thinking this is likely what my hardware guy intended, but I can't raise him atm
Our needs are really modest (only need i2c and BLE, really)
When I inspect the chip memory, CP takes up about 50% of the available space
tulip sleet
Our needs are really modest (only need i2c and BLE, really)
it depends more on your code size, including the libraries
it's more likely he would ask you that q, and you would spec it
pca10059 and raytac_mdbt50q-rx builds are examples of builds that use INTERNAL_FLASH_FILESYSTEM = 1. There are others as well.
daring pumice
pca10059 and raytac_mdbt50q-rx builds are examples of builds that use INTERNAL_F...
Ok. I'm currently basing my build off of the bluefruit feather sense - wondering if that's wise.
tulip sleet
pick one of the no-external-flash builds instead
daring pumice
And if so, do I need to include INTERNAL_FLASH_FILESYSTEM = 1
tulip sleet
definitely
daring pumice
pick one of the no-external-flash builds instead
Ok. Can you suggest one? Sorry to be a pain, but it's hard for me to distinguish between appropriate/inappropriate base builds
Just don't know enough yet
tulip sleet
the raytac module one is pretty simple. You could look at the boards at circuitpython.org and see which one is closest to yours
daring pumice
the raytac module one is pretty simple. You could look at the boards at circuitp...
Ok, I see two: the db-40 and rx
the rx is a dongle?
daring pumice
Ok. Better fit? The db will have buttons I don't have
tulip sleet
boards/ADM_B_NRF52840_1/mpconfigboard.mk
boards/arduino_nano_33_ble/mpconfigboard.mk
boards/bless_dev_board_multi_sensor/mpconfigboard.mk
boards/bluemicro833/mpconfigboard.mk
boards/bluemicro840/mpconfigboard.mk
boards/electronut_labs_blip/mpconfigboard.mk
boards/electronut_labs_papyr/mpconfigboard.mk
boards/ikigaisense_vita/mpconfigboard.mk
boards/makerdiary_nrf52840_mdk_usb_dongle/mpconfigboard.mk
boards/microbit_v2/mpconfigboard.mk
boards/nice_nano/mpconfigboard.mk
boards/pca10059/mpconfigboard.mk
boards/pca10100/mpconfigboard.mk
boards/raytac_mdbt50q-db-40/mpconfigboard.mk
boards/raytac_mdbt50q-rx/mpconfigboard.mk
boards/simmel/mpconfigboard.mk
boards/sparkfun_nrf52840_mini/mpconfigboard.mk
boards/ssci_isp1807_dev_board/mpconfigboard.mk
boards/ssci_isp1807_micro_board/mpconfigboard.mk
boards/teknikio_bluebird/mpconfigboard.mk
boards/warmbit_bluepixel/mpconfigboard.mk
have INTERNAL_FLASH_FILESYSTEM = 1. Some of these are nRF52833 builds, so need to take those out. Then I'd just pick the ones that are the closest. You will certainly need to change some pin defns.
eliminate microbit_v2, pa10100, bluemicro833, arduino_nano_33_ble off the bat
daring pumice
Much appreciated.
slender iron
I don't think there is a way to measure refresh rate
displayio will cap the refreshes though
but if a lot changes, it won't keep up
mental nexus
I didn’t see a built in way, but thought I might be missing something. I’ve never had much luck with changing the displayio refresh rate and observing any differences (by eye).
slender iron
its been a while since I've been in that code
mental nexus
Sounds like I may need to add a custom hook to count refreshes.
idle owl
@jaunty juniper 🐈⬛ !!
onyx hinge
[I can't idly type on that keypad if it's plugged in]
idle owl
[I can't idly type on that keypad if it's plugged in]
That's why I include fidget mode.
onyx hinge
@idle owl maybe you can PR that to qmk in your free time this week.
idle owl
<@330227457296957440> maybe you can PR that to qmk in your free time this week.
Hah! I'll get right on that 😄
modern wing
Thanks everyone!
mental nexus
Thanks everyone!👋
onyx hinge
👋
slender iron
thanks!
gilded cradle
thanks
proven garnet
Thanks!
lone axle
👋 Thanks everyone
minor plume
Thanks, everyone! 👋
modern wing
What is the date for CP day again, @idle owl
idle owl
August 19, 2022!
modern wing
thanks 🙂
onyx hinge
@tulip sleet @slender iron want to re-convene at the top of the hour?
idle owl
You're welcome!
onyx hinge
I could use a moment to stretch my legs & get a beverage refill
idle owl
<@329766224093249548> <@252717193496756235> want to re-convene at the top of the...
I think that meeting was going to wait - Limor wanted to be a part of it. (Unless I'm forgetting mention of a separate meeting).
onyx hinge
oh okay nevermind then
I may have had divided attention during the internal meeting due to the sense of being behind
idle owl
I did as well, trying to get an Actions run to be in progress. 😄
slender iron
limor can do after 4 eastern if we want to meet today
tulip sleet
i am fine with >4 ET
i thought you were off before then, but you are available, fine by me
slender iron
nope, I gotta get lunch soon but then I'm around until 5 pacific
I just mentioned I wouldn't be watching the baby later
tulip sleet
ok, let's ask her in slack
slender iron
(becca is back from her appt)
lone axle
Here is the notes document for next Monday’s CircuitPython Weekly meeting. It is at the normal time of 11am Pacific / 2pm Eastern here on Discord. Everyone is encouraged to attend! Please add your hug reports and status updates even if you’ll be attending the meeting - it’s super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and we’ll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/1tiCKK3V9-X9_jgHTmTlEjDPfB_z7bjZTNu2kYtM7KBw/edit?usp=sharing
hidden oxide
👍🏼
proven garnet
In some of the release CI workflows, right before building the distributions, there this:
for file in $(find -not -path "./.*" -not -path "./docs*" -name "*.py"); do
sed -i -e "s/0.0.0-auto.0/${{github.event.release.tag_name}}/" $file;
done;
I think that's how __version__ is populated right? So should all workflows have this? python setup.py sdist is becoming python -m build in order to build the wheels/source distributions, but I take it we still want that step in beforehand either way, correct?
Trying to figure out if that needs to be patched into all libraries, or removed from ones that have it.
jaunty juniper
speaking of that, is there a standard way for a module to provide a version that can be autofilled when installing with pip or is it always a separate process like this ?
jaunty juniper
I don't think we provide module.__version__ in every library at present time, some are packages with an empty __init__.py (circup finds it in the other files, but it can't be relied on from python code)
proven garnet
I don't think we provide `module.__version__` in every library at present time, ...
Got it, thanks!
proven garnet
speaking of that, is there a standard way for a module to provide a version that...
setup() and pyproject.toml define the version metadata variable. It can also be dynamic and setuptool_scm defines it using the tag defined in the release, which is what we do.
That's why if you have changes from the tag when you install it, it shows up as dirty
Oh you mean from the code itself though?
jaunty juniper
oh yeah I used that in discotool:
setuptools.setup(
...
use_scm_version={
'write_to': 'discotool/_version.py'
},
that results in something like that when not on a tag:
# coding: utf-8
# file generated by setuptools_scm
# don't change, don't track in version control
version = '0.3.2.dev4+ga5d477a.d20220125'
version_tuple = (0, 3, 2, 'dev4+ga5d477a', 'd20220125')
yeah if you wanted to do:
from module import __version__
but that's a different question, I think the point is that modules on pypi don't use or need or care about __version__ but we could make it more consistent accross the CP libraries
tulip sleet
I would be in favor of that. I would like to be able to ask someone what the__version__ of a library is if they import it in the REPL, for any library.
as a support mechanism
jaunty juniper
yeah absolutely
proven garnet
Sounds like it should get patched then, and all libraries should have __version__ defined in them or their __init__.py right?
I just confirmed that currently BME680 library shows __version__ as 0.0.0-auto.0 so looks like it should be patched in then. That patch to release.yml would support what you're saying, @jaunty juniper and @tulip sleet
In a future library patch
We could just add __version__ == "0.0.0-auto.0" in the appropriate place and voila, I think
jaunty juniper
yeah sounds right
manic glacierBOT
orchid basinBOT
proven garnet
I misunderstood what you meant by "discuss", I was assuming In The Weeds, but yo...
Meant to respond that noon works for me!
idle owl
Meant to respond that noon works for me!
Great, I had forgotten, so this works as a reminder. 😄
orchid basinBOT
slender iron
@gilded cradle it looks like we'll need a new build for the lite s3
gilded cradle
ok
slender iron
actually maybe not
looks like backlight is inverted
which would explain why you can only see it when the display ram isn't inited
@gilded cradle have time to test now? what file extension do you need?
gilded cradle
<@478616919721771019> have time to test now? what file extension do you need?
I can test in a bit
slender iron
k, I'm done at 5
gilded cradle
k, I'm done at 5
I would need the .bin file
slender iron
kk
gilded cradle
I can test in a little bit and I'll let you know if it works
slender iron
👍
proven garnet
I think I asked this before but didn't understand the answer: What's the difference between the firmware adafruit_bus_device (written in C) and the Python library adafruit_bus_device? Is the latter only needed when the former isn't available? Because it doesn't look like the Python library references any code from the C code.
slender iron
@gilded cradle
gilded cradle
<@478616919721771019>
Thanks
gilded cradle
<@478616919721771019>
I'm getting nothing. Not even a CIRCUITPY drive with this one.
gilded cradle
I'll try that
gilded cradle
maybe that doesn't have the bootloader 🙂
It's better. It's just rotated 90 degrees and inverted.
gilded cradle
Trying now
gilded cradle
gah, changed the wrong file
The text is now backwards, but the color is correct and it's still at 90 degrees of what it should be.
slender iron
did the orientation change?
slender iron
I'll give you one more build...
gilded cradle
ok
gilded cradle
I think width and height may need swapping too
slender iron
the dimensions are the same... I think it'll be ok if I get rotation right
gilded cradle
ok
gilded cradle
and I'm out
Ok, no change. We can try tomorrow.
orchid basinBOT
manic glacierBOT
bleak matrix
Hi all 👋 . I may have unwillingly discovered an incompatibility between the UF2 images built for STM32F405 boards and TinyUF2. Images produced both locally and downloaded from http://circuitpython.org won't flash! This appears to be due to the family ID contained within the UF2 images being different from TinyUF2's expectation. C.P. builds use a value of 0x6d0922fa (STM32F407) whilst TinyUF2 is expecting 0x57755A57 (STM32F4). I would be grateful for any prior knowledge / advice / suggestions as where to go from here. Cheers.
jaunty juniper
oh I remember noticing that tinyuf2 doesn't flash CP when looking at adding the UF2 to the builds of the STM32 feather. Unfortunately I don't know anything about that family ID business...
bleak matrix
Thanks @Neradoc.
jaunty juniper
I didn't look into it back then, just made a note to get back to it some day. I would open an issue in the CP repo to have eyes on it
bleak matrix
I'll take that suggestion and raise it soon. Thanks for the reply 🙂
daring pumice
If anyone's up, wondering if there's a list for Flashing LED Statuses.
I only have a single-coloured LED (not a multi-coloured RGB LED) and am getting a repeating pattern of 6 flashes after the initial bootup sequence.
tiny token
What are the benefits of creating an .mpy file for a library, other than potentially smaller footprint? If your library project has multiple .py files can you combine them into one .mpy file to simplify its installation/use? I come from a c#/dll background so I can see benefits of a 'compiled' library but not sure if .mpy is anything close to a .dll.
tiny token
can you put multiple .py files into a single .mpy or is it a one-to-one? my end goal is to simplify the library install and related import statements required to use it. I'm not a fan of all classes/code for a library placed into a single .py file.
tulip sleet
can you put multiple .py files into a single .mpy or is it a one-to-one? my end...
it's 1-to-1
onyx hinge
@tulip sleet @slender iron @idle owl maybe either between 1PM-3PM ET today, or else tomorrow for going through the 8.0 milestone list? I am hanging out with a friend from 4-6PM ET today.
gilded cradle
code is here: https://github.com/tannewt/circuitpython/tree/esp_box_lite
FYI, I tried the web workflow on the esp32-s3 box lite and it totally crashed the device until I installed older firmware and removed the .env file. So I'm thinking it may need to be a new board in CP. I also noticed buttons are different, so maybe there's quite a bit more differences than are obvious.
slender iron
@gilded cradle ok. I’ll ping you when I’m at my desk. Kiddo didn’t sleep well last night so I’m getting a slow start
gilded cradle
<@478616919721771019> ok. I’ll ping you when I’m at my desk. Kiddo didn’t sleep ...
No worries.
proven garnet
notes
idle owl
<@329766224093249548> <@252717193496756235> <@330227457296957440> maybe either b...
I'm available between now-ish and 3pm. Tomorrow I'm busy between 2:30-3:30pET.
gilded cradle
<@478616919721771019> ok. I’ll ping you when I’m at my desk. Kiddo didn’t sleep ...
I grabbed the ESP32-S3 Box (non-lite version) from my office and tried absolute latest CP firmware which works fine so far. I think the lite version doesn't have a touchscreen. Probably has a different display altogether, so that's why it looked strange. Not sure on the WiFi though. Anyways, I'll continue web workflow stuff with the non-lite box.
slender iron
I grabbed the ESP32-S3 Box (non-lite version) from my office and tried absolute ...
great! I'll have to pick up a lite one. let me know if you hit any roadblocks
slender iron
I'm available between now-ish and 3pm. Tomorrow I'm busy between 2:30-3:30pET.
I'm at my desk now
idle owl
Dan appears to be idle at the moment.
slender iron
If anyone's up, wondering if there's a list for Flashing LED Statuses.
I only h...
slow flashes at boot are when you can enter safe mode. faster flashes after that are when you can enable the ble workflow
becca has an appointment this afternoon so my availability will depend on whether the kid is asleep
idle owl
@onyx hinge You still around?
tulip sleet
i can do it now. I have a service person here for the heating system so I may need to step away briefly.
at an unknown time
idle owl
Yeah I think so. Somewhere we can vid and share screens. That worked last time.
slender iron
kk
slender iron