do you still have the 2.3.0 firmware installed?
#circuitpython-dev
tidal kiln
acoustic pollen
yes
tidal kiln
go ahead and switch back to 3.0
i don't think this is a firmware version issue, so might as well move forward with the latest version
tidal kiln
the examples are short
we got these files to play interactively using the REPL last night
they should play the same way when a button is pressed for that guide example
acoustic pollen
yeah, console could play the whole wav but it seems that buttons can only play a short bit
tidal kiln
did you also change the version of lib bundle when you changed the firmware to 2.3.0?
acoustic pollen
no
now im trying to see if i can do the random talking code from the sparky automation box minus the servos
so i got it so that pressing the a button does a random wav, is there a way to do it that if it goes fast enough it plays a random wav?
tidal kiln
if what goes fast enough?
acoustic pollen
the board
idle owl
@gusty kiln Agreed with reducing overhead, and agreed without understanding it, I don't have any idea how to either. It's really not too bad once I get it right. It's the getting it right in the first place bit that I'm still figuring out ๐
acoustic pollen
also, when trying to do it so that sound triggers on a1 being touched, using an alligator clip makes it just go off repeatedly when plugged in via usb. is this intentional, or am i just using a bad alligator clip?
idle owl
@gusty kiln So updated workflow:1. I do a PR. 2. You review/change request/merge. 3. I trigger a Travis build and deal with getting it building. Possibly goto: 1. 4. I do the release.
Does that work for you?
gusty kiln
sounds good.
idle owl
ok rad.
I'm going through everything, finding some that I didn't do that don't have overwrite = true. I added it to one, should I keep adding that?
gusty kiln
yeah, i think so.
idle owl
ok
gusty kiln
If you need to overwrite existing files, add overwrite: true to the deploy section of your .travis.yml.
idle owl
oh ok keen
gusty kiln
seems like it mostly won't come up but if it did that's what we'd want.
prime flower
@gusty kiln I have a weird pypi question/issue as well if you have time
(shakes fist) PYPI
maybe you;ve hit this too, kattni.
gusty kiln
@prime flower dunno if i'll know, but hit me
idle owl
same
prime flower
Released a new version of the IO-Python client, merged to master. Released and tagged.
It was tagged on release (release page), but pypi complains that the commit was not tagged.
idle owl
Does it need to be master? Not sure how you would have released a specific branch, but I'm reading that as it not liking a branch.
Otherwise, no afaik I haven't run into that but I haven't been handling travis builds so far. I picked that step up today.
prime flower
maybe my flow is wrong but I was doing: branch -> pr into master -> release master with new tag
gusty kiln
that seems like the correct flow. i haven't seen that error before.
idle owl
That is correct, @prime flower
prime flower
I had it release, twice. And somehow can't get it releasing without that error again.
idle owl
It's tags: true in mine.
prime flower
changing yml, re-releasing/tagging
idle owl
@slender iron Is the only thing that needs to be added to .travis.yml to force the earlier pylint this line under install? ```install:
- pip install --force-reinstall pylint==1.9.2```
prime flower
@gusty kiln @idle owl welp, that seems to have pushed it to latest on pypi!
thanks!
idle owl
np!
gusty kiln
cool
prime flower
dont remember why I took tags false, but at least I'll spot it next time.
slender iron
@idle owl I think so but haven't done it myself
idle owl
@slender iron It seems like it worked, travis passed pylint after I added it. Thanks
slender iron
good enough for me
idle owl
The only one we'd updated was the Learn repo and it had something different under scripts: in .travis.yml so I wasn't sure if that was part of it or how we handle pylint in the Learn repo because of the weird structure. I assume the latter.
@gusty kiln https://travis-ci.org/adafruit/Adafruit_CircuitPython_BME280 Was this the same issue on the other one that wasn't working?
@slender iron Ok, so the build I just linked says it's logged in as not-Adabot. Where is that coming from? Is it the GitHub Token in travis-ci.org?
slender iron
ya
idle owl
Is the way to fix that to delete the GitHub token from travis-ci.org, and run Adabot?
slender iron
thats one way or make one manually yourself
gusty kiln
definitely the same deal as bmp280
idle owl
Ok, so log in as Adabot, create a token, and then add it to Environment Variables on travis-ci.org, correct?
Hmm now it passes but it's claiming it's not a tagged release. I retagged the current release. Running again.
Same thing. Hmm.
@gusty kiln Now I'm getting this instead. ```store build cache
Skipping a deployment with the releases provider because this is not a tagged commit
Skipping a deployment with the pypi provider because this is not a tagged commit
Done. Your build exited with 0.```
gusty kiln
lookin'
idle owl
I redid the release tag like it says to do in the instructions.
I could delete it and do another new release instead.
main meteor
Maybe it's the commit tag, not the release tag it's complaining about?
idle owl
It's based on release tag as far as I understand it.
You create a release, and it uses that tag.
gusty kiln
it looks like 2.0.4 is good?
idle owl
It passes, but it's not deploying at the end. It's not doing the pip stuff in travis.
gusty kiln
ah, my bad - hrm
idle owl
Yeah it's misleading because if it skips deployment, it doesn't actually fail
upbeat plover
is there a circuitpython for this library?
https://github.com/adafruit/Talkie
idle owl
@upbeat plover Not that I'm aware of.
gusty kiln
there may be weirdness because both tags point to the same commit?
idle owl
Oh. Like I need to create a new commit of some sort?
gusty kiln
it doesn't seem to me like that should be the case, but it also wouldn't surprise me.
googles
idle owl
GitHub at least thinks it's tagged.
But both show up there.
There is a way to create an empty commit.
gusty kiln
this error seems to crop up a fair bit: https://github.com/travis-ci/travis-ci/issues/9437
my guess is that travis's code for detecting a new tagged commit is slightly failure prone around some edge case.
idle owl
Ok I deleted the initial tag.
So there's only 1 on the commit now according to GitHub.
Trying the build again
I did figure out there's a way to create an empty commit on git. So we can try that too if you think it might help.
Same.
gusty kiln
that is actually new information to me. :)
idle owl
Passes build, skips deployment.
gusty kiln
probably worth just trying a fresh commit & bumping the version tag.
i'm trying to figure out how travis determines whether a thing is tagged, but it turns out that modern idiomatic ruby is kind of... inscrutable.
(or maybe it just turns out i don't know ruby.)
idle owl
(ruby can go stuff it.)
Blergh. Ok. I'll do the empty commit thing.
@gusty kiln Wait, so I can't delete a release apparently. Do I create a new one? Or will bumping the tag move it to the new commit?
Looks like it might. I'll try it.
Oh can delete it, have to delete the binaries first.
gusty kiln
@idle owl in general we should probably avoid changing what a tag points at
i came to this conclusion after changing what a tag pointed at the other day
idle owl
I deleted the release.
gusty kiln
and then reading the extremely dire warnings about Never Doing That in the git tag man page.
gusty kiln
i can see why, but it's not something i'd given a whole lot of thought in past.
slender iron
tag numbers are free ๐
idle owl
Yah I know. Just wasn't sure what route to get there.
Ok. New release. New commit. New tag.
gusty kiln
cool, fingers crossed
idle owl
Build autotriggered.
YES
YES YES YES
Seriously, travis can go swim with hungry sharks. ๐
It worked
Logged in as Adabot
gusty kiln
whew. :)
idle owl
Ok so I have so many threads of things I'm trying to do here, but I will add to that getting every lib logged in as Adabot. I need a spreadsheet or something ๐
gusty kiln
sometimes that really is the best way, much as office suites grind my gears
idle owl
Agreed, and I'm useless with spreadsheets, but this is HAPPENING RIGHT NOW.
idle owl
@slender iron I have a question about RTD. The badge says docs are failing. I don't know how to get to the dashboard that tells me exactly what's failing unless I'm the one who imported and added it. What am I missing? The docs are there if you click on it, but the badge says failing. I know there's the thing that shows you the builds, but I can't seem to get to that page unless I'm the one who added it so it shows up under my Projects.
slender iron
its in the lower left menu on rtd
idle owl
ohhh... ok. Thank you.
prime flower
@idle owl if you do find the name of it, lmk. I usually build sphinx locally beofre passing to travis to avoid that
idle owl
It's just "builds" in that lower menu. I build it locally too but this is on stuff that failed all on its own.
Appears there's no way to trigger a build without a new commit being pushed to the repo.
So if a build failed due to inactivity, which this one did, there's no way to trigger a new one without pushing a commit to the repo.
No, I was able to trigger a build on one that I imported and added to RTD.
hmm.
Yah button's not there on a different one.
idle owl
@gusty kiln I haven't had this happen yet: https://travis-ci.org/adafruit/Adafruit_CircuitPython_ADS1x15 PyPI upload failed. failed to deploy
idle owl
Oh. Thank you.
idle owl
L800?
deep sigh
If that's the file that already exists, then I have to retag it. Which means doing the empty commit again and deleting the release and redoing it and etc etc etc.
New tag will be new file name if I'm reading that right.
prime flower
@idle owl I was doing 2.0.x starting with 1. Wound up on tag 2.0.9 when it finally deployed to PyPi
(was about to temporarily throw in the towel if I hit 2.0.10 without a successful deploy for that time and drink some water)
idle owl
Got ya ๐
....
I updated it and it failed on the same file name, it's not using a new tag in the filename.
gusty kiln
hrm
idle owl
Current tag is: 0.5.1```just reading scrollback
gusty kiln
it looks like it's trying to upload a tarball for 0.5.0 despite that being a build for 0.5.1
idle owl
Right.
And that already exists evidently because that was the tag previously.
But it sees the updated tag earlier.
My google fu is failing here. I don't know where to begin searching for this. Restarting the build to see if maybe it takes a minute and it triggered too early. Doubtful but I don't know what else to do.
gusty kiln
yeah, digging a bit
idle owl
failed. as expected.
gusty kiln
ok, so i have a working theory
i think setuptools_scm might grab the first tag (0.5.0) instead of the second one.
running an sdist on its current state builds 0.5.0, at any rate.
idle owl
hmm.
deleted the tag from GitHub.
not sure if that'll matter.
@gusty kiln Nicely done!
That worked
gusty kiln
cool.
yeah, in general it seems like any time there's more than one tag on a given commit, the tooling is going to break somewhere.
idle owl
I thought deleting the release was enough to fix that, because it wasn't showing two tags. But noted for future reference.
I'm actually surprised it worked, I didn't do the empty commit.
Forgot that step.
LOL ๐
I HAVE ARRIVED!
I finally did a git blame and it was me!
idle owl
So pylint is failing on this now in multiple repos. Module 'board' has no 'NEOPIXEL' member, but source is unavailable. Consider adding this module to extension-pkg-whitelist if you want to perform analysis based on run-time introspection of living objects. (c-extension-no-member) It obviously DOES have NEOPIXEL in it, but it's not seeing it. So do I disable it? Do we need to update .pylintrc to either ignore no-member or add something to the whitelist? The description of what happens if you add things to to the whitelist doesn't sound great.
In the other repo, I disabled it because I have no idea what was going on in that section of code. In this one, I do get it but don't understand how I could make it not complain.
This is with pylint 1.9.2 supposedly.
$ pip install --force-reinstall pylint==1.9.2 is in there.
@slender iron I need your thoughts on this one.
I know we don't want to default to simply disabling pylint, but I'm wondering if it's the right thing to do here.
slender iron
ah, its c-extension-no-member? that we could disable
idle owl
In other places it's no-member
slender iron
plain no-member can be useful
idle owl
Using config file /home/travis/build/adafruit/Adafruit_CircuitPython_NeoPixel/.pylintrc
************* Module neopixel_simpletest
E: 11, 9: Module 'board' has no 'NEOPIXEL' member (no-member)```From travis-ci.org
So it's different for me locally
The one I pasted earlier is from me
slender iron
hrm
slender iron
does automock fix it?
idle owl
I thought we took that out?
slender iron
instead of using blinka
idle owl
let me try
We never put board in there though that I was aware of....
I'll try putting it in there.
how does pylint get the automock though? I thought that was sphinx specific
idle owl
Yes apparently.
2.0.1 it says. Oh I know why, this is a new venv. I didn't update the other one.
slender iron
could update and blacklist the c error
idle owl
I wasn't thinking when I installed it in this one.
Then we run into other issues I think. But I'll give it a try.
So, remove the 1.9.2 line from travis.yml, and then add that to the pylintrc disable line?
the error I mean
Ugh. I think it's using 2.0.0 and maybe that extra fancy message was 2.0.1?
Because it's still reading no-member on travis-ci.org
idle owl
Ok this is weird. I forced 2.0.1 install. And it still says (no-member)
Is it asteroid that's giving me extra fancy messages? Because pylint --version gives me a list of things. 2710 kattni@robocrepe:neopixel (venvpypi) [5s pypi 14cf74c]$ pylint --version pylint 2.0.1 astroid 2.0.1 Python 3.6.5 (default, Mar 30 2018, 06:41:53) [GCC 4.2.1 Compatible Apple LLVM 9.0.0 (clang-900.0.39.2)]
No that's up to date too Requirement already satisfied: astroid>=2.0.1
I'm disabling it locally and moving on.
@slender iron I thought having Adafruit-Blinka in requirements.txt handled all the built-ins? It's not handling neopixel-write.
Do I readd it to automock?
Or is it supposed to work.
slender iron
it only handles whats been implemented. ya, mock it or skip the lib all together
idle owl
I don't know how to skip it so I'll mock it.
slender iron
don't pypi it at all
idle owl
Oh.
slender iron
it won't work on cpython anyway
idle owl
I thought you were saying there was some magic to skipping neopixel_write in sphinx.
It won't?
Adafruit-Blinka should cover analogio right?
slender iron
yeah but not neopixel_write
idle owl
No module named 'analogio'```Blinka is in req.txt and I at least think I installed the right things in the venv.
Failing remotely too.
idle owl
No but it was all a mystery to me until now, so that's super helpful.
slender iron
๐
idle owl
Is the plan to eventually implement all of it?
slender iron
ยฏ_(ใ)_/ยฏ
idle owl
Fair enough.
leaden kettle
Hey all! Total newbie here...I'm not sure if this is the place to ask, but it's driving me crazy and I can't seem to find an answer on Google. Is it normal for the CPX to show up as two drives on my computer? It seems to be working ok for now, but it was just unexpected. When I first started using it, it only showed up as one drive, but somewhere along the line it decided to turn into two. CIRCUITPY(E:) and USB Drive (F:)...The extra one is empty, and 0 bytes used 0 bytes free. It may not even matter, but I was just curious. Thanks! ๐
upbeat plover
i would put it back into bootloader and throw uf2 on it again, if that doesnt fix it then in REPL use
import storage
storage.erase_filesystem()
backup everything you want to keep first
opaque thicket
n00b question about opening audio files in CircuitPython. I'm looking at the playing .wav file example here: https://learn.adafruit.com/circuitpython-essentials/circuitpython-audio-out
We open the file by saying "wave_file = open("StreetChicken.wav", "rb")"
Can we get away with using a "with" statement, like this?
with open("StreetChicken.wav", "rb") as filey:
wave_file = filey.read()
Does this sort of call only work with text or is this valid?
(If it's valid, are there any advantages/disadvantages between the two methods?)
opaque thicket
Another question: if I'm importing random, is it the same as importing the standard python random library? Like, could I call something like r = random.uniform(0, sum(w))?
onyx hinge
@opaque thicket only a subset of python's functionality is available. It looks like random.uniform is on the menu, according to https://circuitpython.readthedocs.io/en/3.x/shared-bindings/random/__init__.html
opaque thicket
@onyx hinge - Awesome! Thanks so much!
onyx hinge
It was too much to hope that this would simply work on the first try...
opaque thicket
Oh man - I'm not following with what's happening there...
Also, thanks for the readthedocs link - bummed to get no results for list comprehensions
onyx hinge
@opaque thicket just an experiment to see if CircuitPython can be compiled to run inside the web browser
opaque thicket
lol - that would be bananas!
manic glacierBOT
onyx hinge
oh, crafty. emscripten's stack might grow up OR down. That is tricksy and confuses MICROPY_STACK_CHECK
solar whale
@idle owl You've been doing a great job cranking through the the drivers. I was curious if each sensor is being verified to work via Blinka before it gets pushed out to PyPi? I know there are still some issues with some SPI interfaces like the RFM69 and RfM9x but I'm not sure if any others have issues.
prime flower
@idle owl I can't install Adafruit_CircuitPython_PCA9685 from Pip, was it deployed?
prime flower
adafruit-circuitpython-motor, too.
doesn't look like the build went through
oh...
l9, provider: releases -> - provider: releases
dht isn't on there either, same error.
@idle owl @gusty kiln you may want to incorporate into your workflow: travis-lint (gem install travis-lint), then check if the request made it's way to pypi by running pip3 install lib_name after submitting a release
idle owl
@prime flower There are PRs in to all those libs to fix that. We already added those steps to the workflow, it wasn't added to the GitHub issue.
idle owl
I had issues with gem installing. We're not using travis-lint but we are verifying deployment through Travis before marking it completed.
prime flower
on macOS? with brew?
idle owl
Brew isn't gem. Gem is straight Ruby. I was able to brew install travis but not gem install it.
prime flower
huh, weird. I honestly don't remember how I got gem installed on this computer (long time ago)
idle owl
I brew installed Ruby but it gave me permissions issues trying to gem install anything.
prime flower
do you want ruby installation help with that (adam is a rails dev and i can loop him in?) or nah
idle owl
My workflow is convoluted but it works. So for now, nah.
prime flower
ha, np, going to install a pre-pypi pca on this pi for now
idle owl
It should be taken care of today so you should have it by the end of the day.
prime flower
yay, thanks!
idle owl
@solar whale We're doing it backwards a bit. It gets setup with PyPi then we have them queued to test. And if there's issues we resolve them at that point. So at the moment a lot of it's in flux because we're in the middle of the steps.
Check the guides. That's that last step I think, so if the guide has CPython in it, it's been tested successfully.
solar whale
@idle owl OK -- no problem -- I just wanted to give you a heads up about ones I know have issues. There is an issue in the Blinka repo for the RFM9x. Its actually a problem with the SPI implementation and not meant to be "fixed" in the sensor dirive.
idle owl
Ah ok. Thank you.
solar whale
arrgh -- I forgot the cause/fix for this error for travis -- ```Warning, treated as error:
autodoc: failed to import module 'adafruit_tmp007'; the following exception was raised:
No module named 'micropython'
@solar whale I just ran into that too, its part of removing the autodoc moc import and moving to requirements.txt
(if you're working on what I think you are)
solar whale
I'm just trying to create a new library
do I have to change one of the files from cookiecutter?
@prime flower did you find a workaround?
bronze geyser
I've been able to create an atmel start project, build it in the makefile and start debugging w/ JLink on an itsy bitsy. To do this, I had to adjust the linker script by commenting out the rom statement.... i think the linker script is still not quite right (memory mapping). Does anyone know where i can get hold of a linker script that would work w/ itsy bitsy? the makefile script w/ CP doesn't use .ld files, although i've been able to figure out a few things. My challenge is i am not fluent on linker scripts. (?)
tidal kiln
@solar whale look for autodoc_mock_imports in doc/conf.py
solar whale
@tidal kiln tahnks I think I jsut need autodoc_mock_imports = ["micropython"]
tidal kiln
i'm not an expert, but that's what i remember making it work
it tells it to pretend like it actually has that module
tulip sleet
@bronze geyser we do have .ld files - they are in circuitpython/ports/atmel-samd/boards/*.ld
the one used for itsy is chosen in boards/itsybitsy_mo_express/mpconfigboard.mk
bronze geyser
@tulip sleet Thank you. Off i go...
@tulip sleet i'm finding the atmel start "code builder" to make more sense as i bumble deeper :-).
tulip sleet
it makes sense, but it's hard to keep track of updates to their code, and it doesn't provide a way to download the whole set of libraries
bronze geyser
@merry turret Halber ...oh, that explains how $(LD_FILE) resolves....
prime flower
@solar whale @tidal kiln yuup! I have had an issue when importing "pulseio" though through requirements.txt
bronze geyser
what i like about it is it kinda works with the datasheet to give me a clue on the "right" steps to take.
as a learning tool.
tidal kiln
@tulip sleet what are the source code zips for on the bundle release page?
https://github.com/adafruit/Adafruit_CircuitPython_Bundle/releases
idle owl
@solar whale The plan is for Adafruit Blinka to handle that. I think. And if Adafruit-Blinka is in requirements.txt it will handle the built-ins. However. pulseio isn't implemented in Blinka yet apparently.
tulip sleet
@tidal kiln Those are generated automatically by github. I don't think they're very useful -- it's a snapshot of the bundle repo
idle owl
This is the weird middle-flux I was referring to. Not everything is implemented but we're PyPi-ing all the libs. I assume in preparation for it all being implemented.
solar whale
I think @tidal kiln had the fix I needed
idle owl
@solar whale Yes the automock works, I'm simply telling you how it's supposed to work moving forward. The automock isn't supposed to be needed anymore.
tidal kiln
@tulip sleet they're empty, just folders, no files
solar whale
OK thanks !
idle owl
So if it's not implemented in Blinka, etc, then automock it. If it is, the requirements.txt with Adafruit-Blinka in it will work.
tulip sleet
@tidal kiln cause whoever downloads that has to do a git submodule update --init -recursive
solar whale
ok - I can try Bilinka after I get it submitted the first time...
tulip sleet
just like if you cloned the bundle repo
tidal kiln
@tulip sleet ah. interesting. so it's actually a git repo in the zip. gotcha.
idle owl
@solar whale The link I posted has what Blinka has in it. pulseio isn't on that list.
micropython is though.
tulip sleet
it's a piece of a git repo. it has a .gitmodules but not a .git/
idle owl
@prime flower Your DHT PR failed travis on other files. I've been linting everything up and adding the force install to .travis.yml for pylint. If you don't want to deal with all of that, close the PR, and I'll do DHT next and take care of it.
solar whale
@idle owl thanks -- I "think" I uinderstand
prime flower
@idle owl closed, that'll need an automock for pulseio
idle owl
@prime flower Yep I see that in Travis.
solar whale
Travis -- I know where you live ....๐
idle owl
ooooooh it's about to go down.
I'm getting in line.
@prime flower You're welcome to review this one, I couldn't add you to it. https://github.com/adafruit/Adafruit_CircuitPython_DHT/pull/13
After travis finishes anyway.
solar whale
you can have first dibs
idle owl
Missed a couple of things. Oops.
@solar whale Excellent. There's no evil grinning emoji. Disappointing.
There we go.
solar whale
anyy idea what this is complaining about ```
Warning, treated as error:
/home/travis/build/jerryneedell/Adafruit_CircuitPython_TMP007/adafruit_tmp007.py:docstring of adafruit_tmp007:1:'any' reference target not found: Adafruit_TMP007
Yes.
If you reference something like that in a docstring, you need double backticks around it or it tries to actually look for it to use it.
prime flower
@idle owl looks good from the travis build!
idle owl
@prime flower Yep! Do you want to merge it? Then I can start sorting the next steps.
@solar whale Did that get you sorted?
meager fog
@gentle bronze hi for the nrfutil stuff that dan is putting on pypi, we can update arduino too!
solar whale
@idle owl yes -- on to more issues, but I'm working through them'
idle owl
@solar whale That's exactly how it goes ๐
prime flower
@idle owl (I can't merge), but feel free to mergein if you can, then release/tag and check if it deploys to PyPi
idle owl
@prime flower Please review it on GitHub for me then, so there's a comment thread.
Thank you
prime flower
Ok, added comment/review, didn't do a code-based review bc it looks fine to me
idle owl
Yep that's fine.
solar whale
Halleluja!!
prime flower
@idle owl Successfully installed adafruit-circuitpython-dht-3.2.0
prime flower
Looks good, release from PyPi installation matches tag release #.
prime flower
I just soldered the one on my desk up ๐
idle owl
It was started but was missing some things so it should go quickly.
idle owl
@gusty kiln Can you hit this one next? It's a dependency for another one that's already in process. https://github.com/adafruit/Adafruit_CircuitPython_PCA9685/pull/15
gusty kiln
yeah, just saw that dependency - lookin'
idle owl
Thanks
idle owl
@prime flower PCA9685 is set on our end.
gusty kiln
@idle owl restarted build on featherwing
idle owl
@gusty kiln There's another requirement that I apparently missed adding to requirements.txt so that one's on hold for now.
gusty kiln
right on.
acoustic pollen
so is the circuit playground express limited to only importing 6 libraries?
idle owl
Depends on the library.
It's limited on memory. So if the libs are big, you're going to fit less. If they're small, more.
acoustic pollen
so this code is fine for the cpx to run:
import time import random import board import os import audioio from adafruit_circuitplayground.express import cpx
but adding import touchio makes it not work
idle owl
That's because the cpx lib has touchio built in. So you're adding it twice.
If you want to do touchio yourself, don't use the adafruit_circuitplayground.express library.
acoustic pollen
oh, guess i can do my code without touchio then
idle owl
@raven canopy FYI I've started adding the line manually to .travis.yml as I go through and update them. So, doing it manually may not be as big of a deal if we're doing in tandem with something else already happening. I assume your patch checks to see if it's already there. Also I hope I'm doing it right, it seems to be working, but that would figure that it would work but not be right.
@acoustic pollen Touch is built in to the cpx lib. To use touch, use cpx.touch_A1 etc to use the touch pads A1-A7.
acoustic pollen
got that working now
idle owl
The point of that lib is to make everything on the board super easy to use, but it means you don't have as much refined control over each feature because the lib does the setup for you. At the point that you're wanting to do all of that yourself, you move on from using the cpx lib and do it all yourself.
acoustic pollen
also, when using alligator clips on the cpx, capacitive touch just goes rapid fire instead of doing it only on press.
unless air's a conductive material or i need to put it in something first
idle owl
It calibrates on startup so when you add something to the touch pad, you need to restart the board.
Then it calibrates with whatever's on it.
idle owl
You can also set touch threshold if it's too responsive.
lone sandal
prime flower
@idle owl Could you bump register up the list of PyPi to-dos? PCA requires it (https://github.com/adafruit/Adafruit_CircuitPython_Register). Thanks!
idle owl
@prime flower register was already done but there had been another commit merged since the release, so it was skipping deployment due to a lack of a tagged commit. I fixed the docs badge, and did a release. I'm checking it now.
Should be good.
prime flower
@idle owl looks good so far, just installed it and pca isn't complaining when imported
Thanks for the speedy addition!
idle owl
There wasn't much to do.
idle owl
@gusty kiln Sorry about the flurry of emails on merged PRs, I realised I'd forgotten to reference the original issue, so I went back through what I could remember and referenced it.
gusty kiln
right on
errant grail
Help, please: Just updated the bootloader on an old Feather M0 AdaLogger board -- it worked! Is there an AdaLogger .uf2 file out there somewhere?
idle owl
@errant grail It's part of the standard CP releases...
So go to releases and download it. Unless you're asking for something other than what I think you're asking.
errant grail
I found the .bin file out there, but I was hoping to be able to use the most recent bootloader (more convenient than moving the .bin file via bossac).
idle owl
Oh I see. I missed that it was a .bin, sorry
I'm not sure. I guess I thought it had a UF2 in all that. You could build one maybe?
errant grail
I'm willing to try -- back to the steep part of the learning curve! ๐
idle owl
@tulip sleet Does building CP for the Feather M0 Adalogger create a UF2 as well as the other firmware files? Do you know off hand?
Before I send @errant grail on a goose chase.
tulip sleet
checking...
errant grail
@idle owl that's okay, I could probably use the (mental) exercise.
tulip sleet
@idle owl @errant grail the .uf2 does get made, but we don't bother to put it in the release, since it doesn't have a uf2 bootloader
idle owl
@errant grail Yeah but if it doesn't even build it, I didn't want to send you after something that wouldn't happen anyway. There you go!
@tulip sleet Thank you for checking
tulip sleet
there is a uf2conv.py file in the source tree
idle owl
right on
errant grail
Thanks Dan. I probably have the only M0 Adalogger with the uf2 bootloader...
updated it with the .ino file - easy peasy.
humble plover
I'm new to circuitpython but not to coding. I'm happily working through the samples using the mu editor. Very impressed with what my playground express can do. I'm wondering what people are doing for a development workflow. I find the load/save cycle very odd as I'd llike my code on my hard dive in project folders and not on my board (CIRCUITPY) folder which occasionally gets corrupted. Also an observation from a newbie - the REPL/ctrl D thing was confusing. I think there were times (early on) when I was in REPL mode and thought my board was not responsing to changes to my saves in the editor.
errant grail
@tulip sleet @idle owl Found the .uf2 file in the tree. Thanks!
... and it worked! Yippee!
gusty kiln
@humble plover FWIW, on projects i've done, i've just gotten in the habit of writing a tiny makefile to copy code across. might be a good fit for you.
e.g. something like this: https://code.p1k3.com/gitea/brennen/wzrd/src/branch/master/Makefile
humble plover
@gusty kiln thanks - that makes sense. I thought I was missing something. So you are using mu editor but against a folder on your filesystem and then have a terminal to run your make file
errant grail
gusty kiln
i'm using vim, with a hotkey to run make, but i imagine you could do something similar with most editors. (this is where i confess i haven't really tried mu.)
idle owl
@errant grail Nice job!
errant grail
Couldn't have done it without you and Dan (the community)!
idle owl
@humble plover Are you new to Python? Or do you already have a preferred editor that you use for programming?
humble plover
Fairly new to python - I'd probably use visual studio code for this sort of thing.
idle owl
Ah haven't used that enough to know how well it works with CircuitPython.
Which is to say at all.
Mu should be able to work with files not on the board, it'll be a couple of extra steps though because it's designed to streamline working with CircuitPython in the live-on-the-board workflow.
humble plover
I like mu's serial output and I like the 'check' linter button.
idle owl
Yeah it's a great editor ๐
humble plover
but would prefer to use another editor
idle owl
What OS are you running?
idle owl
I use PyCharm. Which has a Terminal built in. And you can use screen to connect to the serial console within the editor.
I learned that recently. I've been using a terminal program and PyCharm side by side.
humble plover
I've got PyCharm installed for when I was playing with tensorflow. It works pretty well then? I thought it might get a bit confused with the cut down syntax
gusty kiln
circuitpython should always be a proper subset of regular old python, i think, so you shouldn't really have any issues there.
idle owl
I've run into a few things. Where it'll complain about libs not being there because it can't find them in CircuitPython, but I ignore it.
It's basically great. I love it.
If you're plugging in a bunch of boards at the same time, you have to do a couple of workarounds to get it to play nice with all of them at once.
humble plover
@idle owl - what arguments are you using for screen to connect to playground express. I thought screen was like tmux, i.e. splitting windows, etc
humble plover
no - just the one board
gusty kiln
@humble plover sudo screen /dev/ttyACM0
should be right for a debian-derived linux like ubuntu
idle owl
Oh right, you have to sudo unless you're in the dialout group
gusty kiln
you might not need the sudo if your user is in the dial... yeah.
this is the one reason i still wind up using screen these days. otherwise it's tmux all the way...
idle owl
Someone this past weekend at the conference tried to connect using tmux, I never heard that they succeeded.
humble plover
Yep that worked! thankyou (ttyACM0)
idle owl
Great!
gusty kiln
yeah, unless something has changed, tmux doesn't do arbitrary serial connections.
tulip sleet
vscode is good -- does an fsync after writing the file, which will ensure it's all written
humble plover
Trying pycharm - I've got the tools > micropython > MicroPython REPL working. Is it possible to set up circuitpython libraries so that I don't get angry squiggles under the imports?
idle owl
Good job with the REPL. I've had all kinds of problems getting it working. I find it'll work once and then not work next time, without me deleting all my settings and starting over which I Finally gave up on doing.
It's not designed to work with CircuitPython per se.
And not really, with the imports thing. That's what I was talking about earlier.
It's the nature of CircuitPython, we have steps all over to deal with it.
humble plover
so no intellisense ๐ฆ
idle owl
PyCharm isn't designed specifically to work with CircuitPython either, so it doesn't know how to look for the libs.
CP is very CPython compatible, but also has it's unique quirks.
humble plover
Well thanks all for the information and tips - very useful! I can get started on a proper project now!
idle owl
Good luck! Keep us posted, we love to hear about projects ๐
humble plover
๐
idle owl
@tidal kiln You around?
tidal kiln
@idle owl am now. waddup?
idle owl
@tidal kiln https://travis-ci.org/adafruit/Adafruit_CircuitPython_OneWire/builds/411934335#L157 Is this error legitimate or is it eligible for pylint: disable
tidal kiln
that link may not have gone to correct place. scrolling down......this?
E: 75,19: Module 'busio' has no 'OneWire' member (no-member)
tidal kiln
np. let me look....
idle owl
Thank you
tidal kiln
why would it say that?
idle owl
Because it can't actually read busio so it's incorrectly reporting it doesn't have that in it.
Or because it's not implemented in the new setup?
Either way, sounds like a disable situation.
I wanted to make sure I wasn't missing something.
tidal kiln
there's no actual issue as far as I can tell:
Adafruit CircuitPython 3.0.0 on 2018-07-09; Adafruit Trinket M0 with samd21e18
>>> import busio
>>> dir(busio)
['__name__', 'I2C', 'SPI', 'OneWire', 'UART']
>>> import board
>>> ow = busio.OneWire(board.D2)
>>>
all as expected
idle owl
I get a fancier error locally: adafruit_onewire/bus.py:75:19: I1101: Module 'busio' has no 'OneWire' member, but source is unavailable. Consider adding this module to extension-pkg-whitelist if you want to perform analysis based on run-time introspection of living objects. (c-extension-no-member)
But we can't get it to find that one to disable it globally, so we are doing it as it comes up.
Ok keen. I'll disable it then.
tidal kiln
but i'm also behind on the latest happenings. whats the "new setup"?
idle owl
Doing all the PyPi stuff. I think it looks in a different place now.
tidal kiln
is it possible to fix the setup somehow?
idle owl
I'm not sure. Because it wouldn't explain why I'm also getting the error locally, except fancier. We could do the thing suggested in my fancy error, but there's a rather awkward sounding explanation of what happens when you do that in .pylintrc so that seemed like a bad idea... something about python "running arbitrary code"
I feel like it has to be added to Adafruit-Blinka or something, but I'm not sure pylint is using that, I only know that it installs the requirements before running pylint.
Something changed somewhere though because this is happening on files I didn't change when I submit the PyPi setup stuff.
I've been checking it and then disabling it if it's definitely invalid.
I don't really understand all of this, I'm learning as I go here.
tidal kiln
you probably know more than i do though.
just wondering if there's some magic something to put in .pylintrc?
idle owl
@gusty kiln If you could do this one up before the others, it's a requirement for the next 5 or something. https://github.com/adafruit/Adafruit_CircuitPython_OneWire/pull/8 no rush, whenever you get to them.
@tidal kiln There is. but I don't understand what it does when you put it there.
@tidal kiln ```# A comma-separated list of package or module names from where C extensions may
be loaded. Extensions are loading into the active Python interpreter and may
run arbitrary code
extension-pkg-whitelist=```
gusty kiln
@idle owl right on - i'll have a run through what's currently in the, and then i think i'm going to call it an afternoon and go stare at my garden for a while. :)
"currently in the queue", i meant to say.
idle owl
@gusty kiln Sounds lovely ๐
I read it right ๐
@tidal kiln I tried to get remote pylint to return the fancy error, including forcing updating it and so on, and I couldn't get it to do it. Kept returning only no-member which we don't want to globally disable. We could globally disable the c-extension-no-member or whatever it was, but I couldn't get it to show up in Travis-pylint. Not sure why. Some kind of dependency of Pylint that isn't updated I'm sure.
tidal kiln
@idle owl what happens if you add busio to that list and run it locally?
idle owl
I didn't try. I'll try it the next time it comes up.
tidal kiln
that first message makes a little more sense to me (i think). since busio is core, it's in C, so pylint can't really inspect it, at the source code level, like a module written in python, so...
Consider adding this module to extension-pkg-whitelist if you want to perform analysis based on run-time introspection of living objects.
would tell pylint to just import it and look at it that way.
idle owl
It is bound to come up again, it has been. I'll try it when it does.
solar whale
@tulip sleet I am running into this repeatedly with the current master, Is something broken ```RuntimeError: maximum recursion depth exceeded
the stack is too small: see https://github.com/adafruit/circuitpython/issues/1085
solar whale
@tulip sleet nevermind -- reading the isse
hmm -- I tried notros fix, but it did not work
tulip sleet
try making it bigger, like 4096 or 8192 instead of 2048 in the line he changed
that only works once, on the next reboot
solar whale
can it be done fron the REPL? or does that have to be done somewher internal
or change it in mpconfigport.h and rebuild
tulip sleet
easier to change it here: https://github.com/adafruit/circuitpython/blob/master/ports/atmel-samd/mpconfigport.h#L343
tulip sleet
@solar whale the corresponding code hasn't been implemented in esp8266
solar whale
Thanks
tulip sleet
could you try importing some large .py file, to see if the stack is big enough for the compiler when it's 8192?
I'm preparing a PR to bump it up to something comfortable for now
solar whale
on teh M4 I use the .py libraries and have not had problems with the RGB display and bitmapfont .pys
tulip sleet
oh, that sounds good, that' what you just tried?
tulip sleet
ok, I'll make the PR 8192 for now -- we can try tuning it later. Also I'll change that error message.
solar whale
Thanks --- Must be Friday night ๐
solar whale
I'm trying to catch up on my todo list -- I'll be pretty scarce for the next two weeks after Sunday -- Going to visit my new grand-daughter ๐
tulip sleet
@solar whale very nice! first grandchild?
solar whale
Second -- our job wil be to entertain the 3 year old!
tulip sleet
congrats!
solar whale
Thanks!
tulip sleet
8k might be high
Adafruit CircuitPython 4.0.0-alpha-919-g177515a7b on 2018-08-03; Adafruit Metro M0 Express with samd21g18
>>> import gc
>>> gc.mem_free()
16240
>>> gc.collect()
>>> gc.mem_free()
16192
Adafruit CircuitPython 3.0.0 on 2018-07-09; Adafruit Metro M0 Express with samd21g18
>>> import gc
>>> gc.mem_free()
19696
>>> gc.collect()
>>> gc.mem_free()
19648
>>>
top is heap size with 8k limit , bottom is 3.0.0 heap size
could shave 3.5k off maybe
solar whale
hmm - I can try with 4K and see how it works
tulip sleet
if you have time, thanks, sounds like you have a good stress test
solar whale
Just take a few minutes
tulip sleet
actually scott put this comment in:
// TODO(tannewt): Make this 6k+ for any non-express M4 boards because they cache sectors on the
// stack.
oh, never mind, that's m4 non-express (of which we have none right now)
solar whale
but I'm using a feather_m4_express and it was failing.
tulip sleet
no, i mean 4k is probably ok, i thought maybe we should make it 6k for m0 basic, but that's not what he means (I missed the "m4")
solar whale
all my tests work at 4K as well
tulip sleet
ok, I'll change my PR. double thanks
solar whale
glad to help.
solar whale
just ran into an odd problem -- I mounted an SD card - then , as I have many times before, tried to use another SPI device, but I get an error that SCK is busy ??
I'll look into this more in the morning and file an issue if need be.
ruh roh -- there were changes to the SPI pin handling 3 days ago ....
Tomorrow is another day to break things -- good night!
manic glacierBOT
tulip sleet
@solar whale do you know when the #1088 problem was introduced? like maybe go back to end of july and e7ae5a3070a1422cf7287f546004b12eec9d0d49. Have you done a bisect before?
solar whale
I have done a bisect -- not overly successfull, but I'll see if I can track it down -- most of the time I have been using the nrf52 board so I'm noat at all sure when this began.
tulip sleet
thanks. it could be the new pin stuff or it could be vfsfat changes from the micropython merge
solar whale
OK -- I see that the new pin stuff is not in the nrf port, is the vfsfat stuff in it?
tulip sleet
yes. I'm actually surprised it worked before, because you're creating two spi objects that use the same pins, if i understand your code correctly.
solar whale
yes, but how could you execute code from an SDCard if you don't
tulip sleet
pass a single spi object around
solar whale
sure, but then you have to have different code for use from SDCard thna from Flash..
tulip sleet
but i don't understand why it worked before. and you say this still works on nrf52 with no error, so it's less likely to the be vfsfat code, since that's shared (or it should be)
solar whale
yes - just ran it on the nrf52
tulip sleet
it's like there's a claim_pin() that was not there before
so the essence is:
spi1 = busio.SPI(board.SCK, board.MOSI, board.MISO)
spi2 = busio.SPI(board.SCK, board.MOSI, board.MISO)
does that work on nrf52 but not m4?
solar whale
really 2 questions -- Yes - I think so
Ill check
does not work on feather_m4_express ```Adafruit CircuitPython 4.0.0-alpha-918-gdfa2581ff-dirty on 2018-08-03; Adafruit Feather M4 Express with samd51j19
import busio
import board
spi1 = busio.SPI(board.SCK, board.MOSI, board.MISO)
spi2 = busio.SPI(board.SCK, board.MOSI, board.MISO)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
ValueError: SCK in use
on on feather feather2832 ```Adafruit CircuitPython 4.0.0-alpha-918-gdfa2581ff-dirty on 2018-08-04; Bluefruit nRF52 Feather with nRF52832
import busio
import board
spi1 = busio.SPI(board.SCK,board.MOSI,board.MISO)
spi2 = busio.SPI(board.SCK,board.MOSI,board.MISO)
Hah!
solar whale
I'll add taht to the issue
tulip sleet
if you tried the same thing with I2C, I think you'd get complaints on both
solar whale
note: I had to type it in manually on the nrf -- cut/paste caued it to hang ??
tulip sleet
that's certainly another bug
solar whale
sigh..
tulip sleet
it's alpha ๐
manic glacierBOT
tulip sleet
then nrf52 is wrong, because the pins should be claimned. Scott's recent changes to that code are only cosmetic. I believe this may have never actually worked on any atmel board -- maybe you are trying things in a new order?
solar whale
I'm not sure I ever tried it on the ATMEl since I don't need the SDCard for the code . On the feather52832, I need it because it wont fit on the intenal Flash.
tulip sleet
the SPI (or I2C or whatever bus) pins are owned by the busio.whatever object, and you have to pass the object around
solar whale
so I need a master script taht mounts the SDCard then execites somethig from it. awkward
tulip sleet
it could be the bare object or it could be a bus_device object
or use a diff SPI bus
solar whale
all workable, just got spoiled by the nrf52 ๐
tulip sleet
you have to be your own OS ๐
solar whale
so any user has to take this intoconsideration when using an SDCard to store code. If the code uses SPI, you have to manage it .
tulip sleet
with an OS, the OS would take care of managing the shared resources
solar whale
I guess this has not gotten much exercise.
tulip sleet
i will open a new bug about pin ownership for nrf52, and close yours in favor of that
solar whale
Is the problem that the 2 instances can conflict with eachother -- so if ai actualy tried writing to the SDCARD while suing an SPI sensor -- things could get ugly.
I feel like I've ratted on the nrf52 ๐
tulip sleet
yes, exactly
or, a worse example, since there's no pin ownership, you could create an SPI bus and and I2C bus on the same pins, and it would not catch that
solar whale
good to find this -- I thik the only way the feather52832 is really nice with an SDCard attached. very limited without it.
tulip sleet
there's a 52840 express board coming down the pike
solar whale
That'll be nice!
tulip sleet
52832 is like feather m0 basic
solar whale
yup -- but more RAM
solar whale
so -- to do thsi properly -- I need to creat the spi instance -- mount the SD Card, then import the libraties I need from it and run my code,.
so my first script -- not on the SDCard has to know waht I am planning to do after I mount the SDCard...
tulip sleet
yes. you're the only person I know who's used an SDcard to hold code to import
solar whale
likely the last ๐
tulip sleet
i think you might have to use a global
because there's no way to pass params to the import
solar whale
It was so nice - I have th whole Bundle on the SDCard -- just mount it and use it like normal
tulip sleet
you could mount it in boot.py and have a global SPI object, or, can you use other pins?
are you using a logger wing?
solar whale
usually either a breakout or the tft featherwing
mostly the tft featherwing.
how do you use the global -- not sure I've done that?
don't take a lot of time -- I can google.
tulip sleet
trying to think of the cleanest way to do this
solar whale
Don't spend any time on ti-- as you said, I'm the only one doing it... I'll see what I can do,
Thanks for helping me understand the problem.
solar whale
thanks . This has taken up much more of your time than I intended!
tulip sleet
no prob - you are always finding valuable stuff! I'll go exercise soon
heavy rain is imminent
for you too - looking at radar map
solar whale
we had some excitement last night! Tornado warnings and nearby lighnting strikes.
tulip sleet
we missed that -- it went north, and you got it. western ma but then it went northeast fast
solar whale
I'm looking forward to Winter !
solar whale
we don't -- its a lot easier t o get warm than to get cool!
its the humidity that is really annoying.
tulip sleet
but no static troubles! - ttyl, thanks for all the testing as usual!
solar whale
You're welcome -- thanks for the fixes --- enjoy the day!
solar whale
@tulip sleet FYI - I now have an M0 adalogger set up so I can mount the SD Card then configure and use my STMPE610 TFT driver via SPI with the STMPE610 driver residing on the SD card. Nothing fancy just had to import the driver after mounting the SDcard and use the same spi instance for both -- here is the example ```jerryneedell@Ubuntu-Macmini:~/projects/feather_adalogger$ cat stmpe610_sd.py
import adafruit_sdcard
import busio
import digitalio
import board
import storage
import sys
Connect to the card and mount the filesystem.
spi = busio.SPI(board.SCK, board.MOSI, board.MISO)
sd_cs = digitalio.DigitalInOut(board.SD_CS)
sdcard = adafruit_sdcard.SDCard(spi, sd_cs)
vfs = storage.VfsFat(sdcard)
storage.mount(vfs, "/sd")
sys.path.append("/sd")
sys.path.append("/sd/lib")
from adafruit_stmpe610 import Adafruit_STMPE610_SPI
cs = digitalio.DigitalInOut(board.D6)
st=Adafruit_STMPE610_SPI(spi,cs)
print("Go Ahead - Touch the Screen - Make My Day!")
while True:
if not st.buffer_empty:
print(st.read_data())
@solar whale lol, nice print.
meager fog
@gentle bronze hi thach - do we have any examples of creating a timer-called function in arduino for nrf52832?
idle owl
Hmm. Getting a pylint error on the fingerprint sensor example, I think I know how to fix it, but I don't have one to test to make sure it works.
idle owl
I have the updated version, but even 1.9.2 is finding new errors for some reason. I'll try it again. give me a sec
Hmm ok passes with earlier version. So in theory it should pass remotely. ๐
gentle bronze
@meager fog yeah, there is a helper class to make use of soft timer of freeRTOS called SoftwareTimer. It is used in the bleuart.ino https://github.com/adafruit/Adafruit_nRF52_Arduino/blob/master/libraries/Bluefruit52Lib/examples/Peripheral/bleuart/bleuart.ino#L22
idle owl
oh wait
meager fog
@gentle bronze i want to get this working on nrf52, but i need to call the LED function at abotu 1KHz https://www.adafruit.com/product/2026
idle owl
no it doesn't pass with 1.9.2
gentle bronze
It is not a obvious example, I will add one more example just for the Software example
idle owl
I added the disable and it autosaved before I ran pylint again
meager fog
@gentle bronze will that timer be consistant or is it a little flexible on timing? ๐
gentle bronze
it is based on the freeRTOS ticks, pretty accurate, but flexible, since other task may not give up control when the time is up
meager fog
ok - i may ask for help later once i get it mostly working
it isnt a big project for me, would be nice to get working but i only know SAMD21 timers ๐
gentle bronze
@meager fog 1 Khz is 1 ms tick, which is too fast since freeRTOS tick is also around 1 khz. you may need to use hw timer for that.
meager fog
can i use HW timer or does FreeRTOS use all of them? yes they need to be fast because it refreshes the whole screen every time!
gentle bronze
@meager fog freeRTOS only uses RTC1 as tick timer
you can use systick
it is easiest to configure
meager fog
ok cool. i hope the timers are easier to use than with nrf51! ๐
ok cool i will do that!
@idle owl did ya resolve or..?
gentle bronze
just a note, RTC is used instead of systick for freeRTOS, since it consume less power
meager fog
ok got it!
gentle bronze
RTC tick still run when cpu enter sleep mode, which systick won't
idle owl
@meager fog It's not passing, and I don't have one to test so I don't want to start refactoring it without being able to test it. I'm going to add a pylint: disable and a note to refactor it when it can be tested, I guess?
meager fog
@gentle bronze i emailed raytac and they are finishing certification, so we'll get final modules in maybe 4 weeks
@idle owl can u link me the error
@gentle bronze so its all very good timing for us. we will have some code ready for hardware. i just need to finish the feather design - that's on me
gentle bronze
@meager fog yeah, thanks for the info, by that time we should have both Arduino and CP ready ๐
idle owl
@meager fog examples/fingerprint_simpletest.py:64:4: R1705: Unnecessary "else" after "return" (no-else-return)
I didn't push it so it's not running on Travis yet. That's local.
This is the block: python print("Searching...", end="") i = finger.finger_fast_search() if i == adafruit_fingerprint.OK: print("Found fingerprint!") return True else: if i == adafruit_fingerprint.NOTFOUND: print("No match found") else: print("Other error") return False
line 64 is the first if
meager fog
@gentle bronze that reminds me to ask, can we reuse the tinyusb code for nrf52840 arduino?
@idle owl ooh yeah, i guess that's true. not my style but correct. how about stick a disable in for now
idle owl
Ok.
meager fog
and we'll look at it later!
gentle bronze
@meager fog yeah, we will need that for CDC + MSC on Arduino code. May need to write some C++ wrapper for tinyusb as well.
meager fog
ok - hopefully peripherals for 840 are similar to 832
we can reuse same bootloader and everything
gentle bronze
I am actually focusing on the Arduino side now, there is quite some works to get it ready especially with the usb as you said, new SD v6 API. Making usre most code work with both 832 and 840
meager fog
also, with appveyor making mac and windows binaries, we dont have to have complex install instructions
gentle bronze
yeah, some user may be confusing at first, though we could just tell them to upgrade the BSP to lastest ๐
the prph is pretty much the same, I only tested uart and pwm https://github.com/adafruit/Adafruit_nRF52_Arduino/issues/143
but pretty much the same, do more testing later
meager fog
yep ok let me know when u want to do the update to nrfutil procedure, i can update BSP/test on mac, win7 and win10
gentle bronze
Thanks, I will try it locally first, then ask you test it again when it is all ready ๐
meager fog
๐
gentle bronze
๐
gentle bronze
@meager fog just in case you haven't check the systick, here is the code #include <Arduino.h>
extern "C" {
void SysTick_Handler(void)
{
ledToggle(LED_BLUE);
}
}
void setup() {
// put your setup code here, to run once:
SysTick_Config(F_CPU/1000);
}
void loop() {
// put your main code here, to run repeatedly:
}
meager fog
ok!
gentle bronze
don't forget teh extern "C" , otherwise it won't link with the one in the asm file. Take it 15 minutes of scratching head ๐ฆ
take me ๐ฆ
idle owl
I just had one of those moments with something else, except it was a day and a half of head scratching. ๐
gentle bronze
very easy to forgot ๐ฆ
idle owl
Those are the most frustrating ones ๐
gentle bronze
yeah, really frustrating, no warning or whatsoever. And we just copy the code from a working proejct ( which is pure C ) ๐
I will add systick example to Arduino example as well. It i useful. not-used and much easier to confi that other phrph's timer.
main meteor
I had a somewhat similar moment trying to call a C++ routine from an interrupt. I finally just had a short C routine that set a flag, then C++ would check the flag. It works, but it's a little tricky to generalize.
gentle bronze
@main meteor nice trick, another way I use is declare an extra extern C func in a cpp file. Then call that function from the c file. The extern C func could call all other c++ funcs without any issue.
main meteor
Ah, I see, function chaining. A C function in the library handles the interrupt and calls another C function in the file that uses the library, which in turn calls a C++ function. I'm unsure how much overhead all that would add (especially the C++ call), given that interrupt handlers are supposed to be quick.
gentle bronze
๐ yeah, it has a bit of overhead, kind of annoying.
But we could defer instead of invoking it right away. It helps to address the linkage from C to C++
Sometimes, the overhead is acceptable too ๐
Hmmm, setting a flag is actually a deferred isr method. I guess it is a better approach for a more robust system. If the C func is not isr, chain call is easier to write :)
gentle bronze
๐ค
idle owl
@gentle bronze Night, Thach! Have a good night!
gentle bronze
@idle owl thanks, good night (afternoon)
scarlet fjord
I have a question about speed with checking for a switch press: is it equally as fast to look through an MCP23017 as it is to do matrixing...? or is one faster than the other?
scarlet fjord
like, in terms of latency and clock cycles etc
scarlet fjord
im making a keyboard, so matrixing in that context would be having 'rows' and 'columns' and checking for keypress events that way.
so a key press check would be:
-turn on the row as an output
-turn on the column as an input
-check to see if the key was pressed
-repeat for each key in the matrix
whereas for the MCP23017, its just:
-check the key to see if it was pressed
so im wondering if checking through the MCP23017 takes more clocks than enabling the GPIOs as inputs/outputs/etc for the matrix
im just worried about latency when cycling through an entire keyboard
tulip sleet
@scarlet fjord Are you not going to use a matrix with the MCP23017?
@slender iron built a keyboard as one of the first ever CircuitPython projects. It uses matrix-ing with a shift register to read the columns: https://github.com/tannewt/gh60 https://github.com/tannewt/ckd63 https://imgur.com/gallery/AAtnz#KZ9K3AL
scarlet fjord
I am not at the moment no ๐ I'm just using three MCP23017s
interesting - I'll check those links out ๐ฎ
tulip sleet
anyway, it should be plenty fast either wya. You can run the MCP23017's at a good clip, especially with SPI. But I2C is probably plenty fast enough too.
scarlet fjord
hmm ok...
I've run some testing and found that it's possible for me to skip a key when pressing it relatively fast. would switching to the SPI version of the MCP23S17 make a large difference to the time it takes to check each pin? I have no idea where my bottleneck is unfortunately lol I'm just not that knowledgable about this stuff
it was cool to check out that imgur link though - I love looking at custom keyboards that people make. I'm wondering what the shift register is for though?
tulip sleet
The shift register is to read the column values. There's only one pin used for the columns. The rows each have a separate pin.
The missed keys may have more to do with the speed of the CircuitPython code rather than the raw speed of the MCP23017. Why don't you just measure the latency? Get the time.monotonic() value at the begining and end of the key scan and subtract them to see what the elapsed time is.
scarlet fjord
hmm I could, yes.
I guess I could test two keys on one MCP23017 vs one key each on two MCP23017s and see if the time is different... that would at least tell if I compared those two results how much time it takes to read the MCP23017... hmm
I guess I should actually put together a matrix and do that testing too to see the time there as well and compare it
I guess I was just hoping someone had already done that testing lol ๐
tulip sleet
I thought nearly all keyboards were matrixed, for cost reasons. Are there existing designs that have used MCP23017 or similar for non-matrixed keyboards, or are you striking out on your own?
scarlet fjord
I haven't seen any that use them, I'm just using them cause I figured it would be easier to understand than trying to learn how tf the diodes fix the n-key rollover and how the code for row/columns works
for my code it's just 'yo pls check this key for a press' instead of some trinary gpio magic lol
tldr; I'm lazy
I guess a matrix isn't so hard.. a for loop within a for loop
I'll try that out tomorrow
rotund basin
Hey all , I am trying to pull a pin low on adalogger , but repel is saying no DriveMode attribute or Pull has no attribute down ๐ค
tulip sleet
@rotund basin Could you paste the code you're trying here?
rotund basin
I'll have to login via laptop hold on
idle owl
I just hit a personal record. Your code has been rated at -47.18/10
(It's not my code.)
tulip sleet
this code is so bad it runs backwards
idle owl
Pro tip: don't use wildcard imports. Your code has been rated at 7.69/10 (previous run: -47.18/10, +54.87)
That's all from fixing the wildcard import. The rest is stuff we ignore in .pylintrc.
rotund basin
@tulip sleet This used to work but doesn't def initMotors():
renable = digitalio.DigitalInOut(rightMotorEnable)
renable.switch_to_Output(pull=digitalio.DigitalInOut.Pull.UP)
sorry don't know how to embed code format
tulip sleet
use triple backticks for multiline code
rotund basin
renable = digitalio.DigitalInOut(rightMotorEnable)
renable.switch_to_Output(pull=digitalio.DigitalInOut.Pull.UP) ```did you import digitalio?
rotund basin
import pulseio
import digitalio
rightMotorPin = board.D5
leftMotorPin = board.D9
rightMotorBrake = board.D10
rightMotorDirection = board.D11 ##must brake before change dir)
leftMotorBrake = board.D12
leftMotorDirection = board.D13 ##must brake before change dir)
rightMotorEnable = board.A0
leftMotorEnable = board.D5
###rightMotorTicks = board.D10
###leftMotorTicks = board.D11
def initMotors():
renable = digitalio.DigitalInOut(rightMotorEnable)
renable.switch_to_Output(pull=digitalio.DigitalInOut.Pull.UP) ##low enables motor
lenable = digitalio.DigitalInOut(leftMotorEnable)
lenable.switch_to_Output(pull=digitalio.DigitalInOut.Pull.UP) ##low enables motor
rbrake = digitalio.DigitalInOut(rightMotorBrake)
rbrake.switch_to_Output(pull=digitalio.DigitalInOut.Pull.UP) ##low activates brake
lbrake = digitalio.DigitalInOut(leftMotorBrake)
lbrake.switch_to_Output(pull=digitalio.DigitalInOut.Pull.UP) ##low activates brake
hmm - it's switch_to_output (all lowercase), not switch_to_Output
also
digitalio.Pull.UP not digitalio.DigitalInOut.Pull.UP
rotund basin
ahh
tulip sleet
cause Pull could be used other places besides DigitalInOut
rotund basin
it's cool. thanks Dan.
tulip sleet
yw!
rotund basin
extra keyword argument at renable.switch_to_output(pull=digitalio.Pull.UP) ?
maybe pull= not needed?
tulip sleet
@rotund basin it's not pull=, it's drive_mode= (had to look it up: https://circuitpython.readthedocs.io/en/latest/shared-bindings/digitalio/DigitalInOut.html
pull is for inputs
rotund basin
oh hmm...
tulip sleet
drive_mode is for outputs: https://circuitpython.readthedocs.io/en/latest/shared-bindings/digitalio/DriveMode.html
are these inputs or outputs
rotund basin
enabling motor drivers and doing PWM so they are outputs , have to output Low on some pins
tulip sleet
you mean you are supplying +V externally and you need to sink current? You have a transistor on the pin output?
what are you driving with the pins?
rotund basin
no, i have motor drivers and I just set pins
tulip sleet
what kind of motor drivers?
tulip sleet
_External input adjust: the terminal of external potentiometer connect to the GND and + 5v terminal, connect the regulator terminal to SV, not only make it possible to adjust speed by external potentiometer, but also can achieve speed adjust through other control unit(Such as PLC, Microcontroller, etc) input
analog voltage to SV. The acceptance of SV is DC 0V~+5V, and the corresponding motor rotate speed is 0
to rated speed. You also can use external digital signal to adjust speed: apply PWM with 5V amplitude and
1KHz ~2KHz Frequency between SV and GND to adjust the speed. The motor rotate speed is adjust by the
duty radio liner adjustment. _
so you just apply a 5v PWM to the SV pin. If its duty cycle is zero, speed will be zero, if I understand correctly.
oh, I see, brake signals, hold on
just make the Brake pins normally high, and switch them to low (False) when you want to brake.
you can use DriveMode OPEN_DRAIN, but it probably doesn't really matter. Looks like floating means not braking.
rotund basin
rbrake = digitalio.DigitalInOut(rightMotorBrake)
rbrake.switch_to_Output(drive_mode==digitalio.DigitalInOut.Pull.DOWN) ##low activates brake
lbrake = digitalio.DigitalInOut(leftMotorBrake)
lbrake.switch_to_Output(drive_mode==digitalio.Pull.DOWN) ##low activates brake
no, not a pulldown or pullup. Outputs don't have pulls. The output is either +3.3V or ground (drive_mode=digitalio.DriveMode.PUSH_PULL), or it's no-connection and ground (drive_mode=digitalio.DriveMode.OPEN_DRAIN). https://circuitpython.readthedocs.io/en/latest/shared-bindings/digitalio/DriveMode.html
Note that this is a 5V device you're driving, so using a 3.3V signal for speed and other controls may be marginal. OK for braking pins, since they are just care whether they're grounded or not.
also you are using == in the code sample above; use = for setting keyword args
otherwise you just wrote an expression that's a boolean value (and there's no variable named drive_mode)
rotund basin
Oh dear I think my metro was 5v
Will adjust code when I get back, thanks @tulip sleet
tulip sleet
@rotund basin all the CPy devices are 3.3v, except that itsybitsys have a single 5v output pin (meant for driving neopixels mostly). But here are a number of level shifters available. Looking at the motor controller data sheet, it says "Low -10~0.5V High 3~30V" for the logic pins, so maybe ok. But if you are sending a speed control signal, that will never rise above 3.3v, so you will not get full speed without a level shifter.
rotund basin
It's cool I have some level shifters
rotund basin
weird behavior i get in 3.0 .... ``` >>> drivemotor.rightMotor(100)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "drivemotor.py", line 28, in rightMotor
ValueError: Pin PA15 in use
i am not using PA15
tulip sleet
@rotund basin which board is this, and could you post the whole codefile? Use gist or upload the file with the "+" sign over to the left.
bronze geyser
hmm.. i'm using CP 3 w/ build that includes RFM69. I do import adafruit_rfm69 I get ImportError: no module named 'adafruit_rfm69' what am i missing? (thank you).
tidal kiln
@bronze geyser that's not a built in module. you need to install the library.
bronze geyser
@tidal kiln oh...thank you.
tidal kiln
@bronze geyser sure, no problem. were you able to figure it out and get it working?
manic glacierBOT
turbid radish
Just a friendly plug - if you'd like to get the spam-free Python on Microcontrollers newsletter, go to adafruitdaily.com. 
๐ถ
torpid goblet
@slender iron Not going to be able to make todays meeting. Off earning money....
Hug report: Special thanks to Dan for rebasing off the latest upstream micropython.
That is a thankless task that requires great patience.
Rosie status: Spent most of last week experimenting with development hardware and a
Development environment. My current approach is to rewrite the logic of
https://github.com/adafruit/rosie-ci using the structure and best practices of Miguel
Grinberg's book at https://flaskbook.com/ . Learning git.
idle owl
@errant grail I noticed that same pop, I attributed it to initialising the DAC, seems I was right. It was louder than the tone for me. I didn't think there was any way around it.
errant grail
@idle owl The pop can probably be avoided if the DAC is slowly initialized to the mid-point value, 65535/2. Slowly could mean stepping from the DAC's current value to the mid-point over a span of 30 to 40ms. Longer would be quieter.
idle owl
Interesting
errant grail
Probably doable in the library.
In the recording biz it's called managing the sound signal's zero-crossing point.
idle owl
Fancy
Seems like a reasonable amount of time too. I'm assuming essentially undetectable by the human ear.
errant grail
It may also be useful to precondition the DAC when the audioio library is first called so that when playback is started, the delay would be shorter.
It's in the < 30Hz range. Still audible but not very noticeable since it's a softer slope rather than a sudden pulse.
idle owl
I meant the delay being undetectable. But that's good to know
errant grail
Ah. The DAC is still sending out a sound during the delay, akin to a quarter of a single wavelength -- but at a loudness of up to 65535/2.
manic glacierBOT
errant grail
@idle owl Yes, exactly. Trying to slow down that initial transient.
slender iron
<@&356864093652516868> Here are the meeting notes for the meeting in an hour: https://docs.google.com/document/d/1-bCCpSRXZo84OsLJNklUst2FH8BqO1kotYfjBYwOcR8/edit?usp=sharing
ruby atlas
I'm glad I looked through the PRs. @slender iron you commented in PR https://github.com/adafruit/Adafruit_CircuitPython_DotStar/pull/23 about the use of an extra buffer for brightness=1.0 which is something that will be the default under the pixelbuf/rgbled/dotstar/neopixel work I'm (finally) continuing to do. I need to keep in mind that that there should always be a way to instantiate a neopixel or dotstar with a single buffer (PixelBuf and RGBLED will support that).
manic glacierBOT
solar whale
@slender iron I'll miss this weeks call ( and probably next week, too) Thanks to Dan for explaining and fixing the stack issue. Thanks to Scott for reviewing the tmp007 PR. Group hug to all. Status - finally got tmp007 driver submitted PR in review. Won't be able to test much for the next few weeks. Will concentrate on on reading/learning.
ruby atlas
I've been struggling with what to call the per-pixel brightness for Dotstar. It's not quite alpha, but RGBB is a terrible constant for RGB+Brightness and RGBA isn't quite the right name.
drowsy geyser
I'll be in listen-only mode today (shared work space).
tulip sleet
@ruby atlas "intensity", or is that too much like HSV?
slender iron
<@&356864093652516868> Meeting in four minutes
idle owl
@ruby atlas Do a mic test, you were quiet last week.
stuck elbow
@slender iron it's my pleasure, can't unmute again :(
raven canopy
this laptop's "cinema sound" makes the meeting sound quite interesting...
ruby atlas
having issues with audio output. sec, restarting discord.
i may have to reboot, discord won't output voice to the headset
brb
drowsy geyser
True. My fault. ๐
raven canopy
mazel tav!
raven canopy
Battle Tanks (i think that was the name) was my favorite 2600 game...
tidal kiln
subhunt!
idle owl
@turbid radish Mute until it's time ๐
turbid radish
ty
wraith tiger
Just lurking today.
idle owl
@wraith tiger Ok thanks!
turbid radish
๐ @idle owl & @gusty kiln
wraith tiger
@raven canopy Battlezone actually. In the arcades it used vector graphics as I recall.
gusty kiln
mostly @idle owl :)
turbid radish
It did @wraith tiger
idle owl
๐
raven canopy
@wraith tiger that might be the one... or Combat. 20 odd years ago... ๐คท
drowsy geyser
Hugs to @idle owl for blasting out so many library fixes over the weekend!!!
wraith tiger
Pebkac?
ruby atlas
in discord, wouldn't it be pebmac to unmute?
raven canopy
OneNote pasting fail....
Group hug...i've vacationed more than anticipated this week...
hard to work with this view.. ๐
idle owl
How awful.
onyx hinge
that one's rad @slender iron (qstr)
idle owl
@drowsy geyser Do you have a status update to read out?
turbid radish
(he says he doesn't speak Russian
neat folio
you can skip me no status.. I am not "working" on anything, just playing ๐
onyx hinge
nope
gusty kiln
@idle owl can probably grep for that pylint-disabling string across repos as a check...
idle owl
@gusty kiln I'll follow up with you on that for sure.
gusty kiln
cool
neat folio
intensity?
raven canopy
Again, not much progress this week.
Adabot Patches: Code of Conduct patch was successful on all library repos. The .travis.yml patch for pylint failed across the board. Realized the problem was the diff result during patch file creation. Basically it included a line that was unique in each library, which raised the "not applicable" flag. I've figured out how to make a patch file without the unique lines, and how to run the git am so that the error isn't flagged. Currently working on adding command line arguments to the adabot script so that any necessary flags can be sent through to the am command. Should have a PR in this week.
FrequencyIn: Only took like 2 momentary stabs at correcting the FREQM interaction; nothing to report beyond that.
@idle owl ๐ i figured you would. there was only about 25 repos that it fails travis on...
tidal kiln
GPS only uses uart.write() and uart.readline()
raven canopy
or....
wraith tiger
RGBi maybe
turbid radish
off to another mtg, next week y'all
raven canopy
๐
onyx hinge
RGBL
tidal kiln
RGBฮฑ
onyx hinge
L = luminance
stuck elbow
HSV uses Hue Saturation Value
idle owl
I know there's at least one person who has asked about being able to set the brightness this way for DotStars.
gusty kiln
i wondered about "luminance", but i dunno how that differs semantically from "brightness". might just get kinda confusing.
stuck elbow
I wonder how it actually metches the terms used in color calculations officially -- there is luminance, intensity, value, etc. -- they are all slightly different things
also gamma
slender iron
If anyone else has a topic let me know otherwise we'll wrap up
raven canopy
what about candle?
onyx hinge
@slender iron meme
tidal kiln
@slender iron i do
slender iron
kk
tulip sleet
how about lower case "brightness": GRBb
idle owl
That could prove confusing.
gusty kiln
hehe
neat folio
many of these terms are the historical results of previous official organisations tackling the same problem and "issueing a decree" ๐
errant grail
Luminance?
gusty kiln
"glowiness"
gusty kiln
no better way to nerdsnipe a bunch of nerds than asking them to name a thing.
stuck elbow
sparkliness
errant grail
blind-esque-ness
stuck elbow
power
raven canopy
@ruby atlas ?? https://en.wikipedia.org/wiki/Candlepower
ruby atlas
LLOL
raven canopy
๐คฃ
neat folio
ooooo... reference... http://www.babelcolor.com/index_htm_files/A review of RGB color spaces.pdf
onyx hinge
@slender iron I had an experience a couple weeks ago where on Metro M4 Express, the AnalogOut object would misbehave: the board would freeze when I tried to create a DAC for the A1 pin. This misbehavior would survive clicking the "reset" button, but would go away when I actually removed power from the whole board. What would NOT be reset by using the physical reset button? Only passives (some resistors) were connected to the A0/A1 pins at this time.
neat folio
gamma seems to be the closest/most similar thing to what we are talking about ( although there might be color shift involved there too)
raven canopy
dynamic clocks survive, yes?