@ionic elk are you asking for your own dev environment or something else?
#circuitpython-dev
pastel panther
pastel panther
hey @raven canopy it looks like the most recent adabot CP libraries report is saying that these two libs haven't been released in quite a while when the both have been released in the last month or week:
https://github.com/adafruit/Adafruit_CircuitPython_BNO055 (334 days)
https://github.com/adafruit/Adafruit_CircuitPython_PyPortal (258 days)
Can you take a look?
onyx hinge
despite my own nay-saying I put together an aws lambda for mpy-compiling. I will take it down if there are problems or abuse. https://q41mecz5na.execute-api.us-east-2.amazonaws.com/mpy-cross @tough flax
it uses mpy-cross.static and https://gist.github.com/98dab44c9b21d7a8a759a73686117a3d
in case of errors it seems to just say "internal server error", too bad π
raven canopy
@pastel panther there seem to be some issues with the GitHub API releases endpoint, recently. adabot's daily report has also had issues. if you look at the websites own compare link, it show older commits as occurring after the last release.
pastel panther
eew
raven canopy
hehe
pastel panther
did you check if there's any bugs filed about it?
raven canopy
i imagine the backend on both of their API services is a spaghetti nightmare. i checked the status and blogs friday night. nothing...
gotta run real quick. hmu if anything comes up.
ionic elk
@pastel panther I'm writing a Zephyr guide and wanted to make sure I was following convention
pastel panther
@ionic elk what are the different options you're considering?
ionic elk
Zephyr has a lot of pip installs and the official startup guide recommends virtualenv. I just wanted to make sure there wasn't something out there saying use Conda.
then I guess there's pipenv? Don't know much about that one
pastel panther
eeew conda no
trim elm
@raven canopy, @pastel panther and I were having issues getting circuitpython_libraries.py to run. I imagine this has to do with the issues you were talking about earlier, but we figured we should let you know anyway. I got a Type Error about string indices needing to be integers, then both Siddacious and I got a Runtime Error which was related to a broken connection. Siddacious was getting a Value Error, which I never got. After I stopped getting the Runtime Error, it went back to the Type Error.
pastel panther
I mean I'm sure it works but it seems like overkill
@ionic elk as far as I know anaconda isn't suggested anywhere in guides except for when using Jupyter notebooks
ionic elk
I've used virtualenv before, but I'm not really a python person so I don't have a particular stake. I hear pipenv is newer and/or better?
pastel panther
not sure myself. I bet @ivory yew would have some good insights
ionic elk
@pastel panther thanks for insight anyhow - will stick to virtualenv unless told otherwise.
pastel panther
solid choice
ivory yew
If you're already comfortable with virtualenv just stick with that.
Oh if you're writing guides yeah def stick with virtualenv.
raven canopy
@trim elm yep. kattni and i discussed that on Friday. the releases endpoint was returning the following, which obviously isn't the JSON the script is hoping for. at minimum, the error could be handled better...
{
"message": "Server Error"
}
Ah, ok
wind trellis
@onyx hinge Nice work on the mpycross service! I can report success using Firefox on Linux Mint and a Circuit Playground Express with 5.0.something. Not that I had any doubts. π
onyx hinge
@wind trellis thanks for your report!
solar whale
@onyx hinge just ran the mpy-cross from MacOS (catalina) Safari -- seems to work well -- I did not execute the .mpy, but "what could go wrong..."
onyx hinge
@solar whale hee hee!
plucky flint
Hi folks, a quick announcement: the latest bug-fix stable release of Mu (1.0.3) is here: https://madewith.mu/mu/releases/2020/01/26/mu-1-0-3.html (fixes the rendering bug in OSX Catalina). Sorry it's taken so long, but this is a volunteer effort and it's taken all day to test, prepare and package the release (in addition to the fixes applied over the past month).
plucky flint
Please report any problems you may encounter. Thanks. π
meager fog
@ionic elk hihi
testing meobit
works great
here's a nice demo printing i2csensor data
meowbit sensor
fiery silo
Hey yβall! A question: before I go about potentially reinventing the wheel, Iβm wondering if thereβs currently a CircuitPython library for reading directly from a SPI flash chip. Like not even file system stuff, but rather just reading data from a given address
meager fog
@fiery silo yeah
i think so
well, there's at least https://github.com/adafruit/Adafruit_CircuitPython_FRAM
you could start that
fiery silo
Awesome, thanks! Iβll take a look
ionic elk
@meager fog thanks for testing! taking a quick CPy break to bang out that Zephyr tutorial BTW
river quest
we now have 2 boards that have meow in them: https://circuitpython.org/board/kittenbot_meowbit/
idle owl
<@&356864093652516868> Here is the notes doc for tomorrow's CircuitPython Weekly. Everyone is encouraged to attend! Please add your hug reports and status updates, even if you'll need attending. It's really helpful! Looking forward to seeing you all tomorrow! Check the pinned messages for this link if you miss this message. https://docs.google.com/document/d/1FqFlX9sNJzdI8y41KqMu7kV0GPW8o7mzWxOoowAwOqA/edit?usp=sharing
obsidian compass
Thanks a lot for Mu Release @plucky flint π
plucky flint
@obsidian compass no problem. Please let me know if there are any problems.
orchid basinBOT
idle owl
@timber mango Discord renders Markdown. if a:
use backticks.
One on each side for inline code, three on each side for a codeblock
@lone axle Are you around?
lone axle
@idle owl I am
idle owl
@lone axle You're foamyguy on GitHub as well, correct? Same avatar as here?
lone axle
Yep, that is correct
idle owl
Alright, I sent you an invite to join our review team.
And I'd like to discuss what that means.
@lone axle This means you have edit/write access to all of the CircuitPython library repositories, which means you have merge access. Under no circumstances should you merge your own PRs. As for merging others, until you're comfortable with it, you can still involve me in verifying that a PR is ready to go, but eventually, you'll get to a point where you're comfortable with it and you can begin merging without a second set of eyes on something. Please be aware that you can now make changes to these repos, so pay attention to what you're clicking on, as the safety net of read-only is now gone. π
This should also mean you that "Edit" link should be available to you now to edit the open PR we were discussing.
lone axle
Okay, yes I will definitely continue to involve you for now until I get a bit more comfortable with it. The great responsibility along with this great power is noted. I will be very careful with github buttons.
idle owl
I'm suggesting that based on personal experience π I've clicked merge when I didn't mean to before. It's version control, so it's always fixable, but annoying when needed.
lone axle
For sure π I will get the one we discussed previously edited later on today. Thank you
idle owl
Thanks so much for everything you're doing! I really appreciate it. Let me know if you have questions when you get to editing, the GitHub GUI is pretty straight forward, but it's not local editing, so it can still be odd.
lone axle
Will do.
timber mango
@idle owl Thanks.
@gilded cradle Good morning. Did you get my email re: changes to my PR?
raven canopy
@pastel panther I just re-looked at the BNO055 release thing. It's not API related. The last PR merged was just an old one. The infrastructure check is using the commit dates, so that's why it shows that it hasn't been released in so long.
pastel panther
@raven canopy ah! good to know, thanks
idle owl
@gilded cradle Please let me know when you're around this morning. Thanks!
raven canopy
@idle owl also, to follow up: it appears that the core repo might be the only one experiencing release API problems. All my other random tests work...
onyx hinge
@lone axle hey I see you are taking on a bigger role in reviewing. thank you!
idle owl
Because the lib report is failing consistently now. The data for the meeting is old.
raven canopy
Yeah. I should've given it more attention this weekend. It's my priority this evening.
obsidian compass
My Python on Hardware poster for PyCon US 2020 has been selected π 
obsidian compass
Thank You! @idle owl
raven canopy
this is why "status meetings" work. forces us to verify information before presenting. π
idle owl
@raven canopy Do you happen to have lib/Blinka data that's more recent from your tests? Not sure what you were testing.
raven canopy
No. I was just calling the endpoints in chrome, with different repos.
raven canopy
If my cell signal would cooperate, I could probably get a useable one...
idle owl
Cooperative cell signals are a myth.
timber mango
I depend on my cell for internet, so my signal better cooperate.
lone axle
@onyx hinge I'm happy to help out where I can π Thank you.
timber mango
@gilded cradle I believe I have it now. PyLint likes me. π
gilded cradle
Yeah, I'm re-reviewing it now π
half sedge
I am looking for a list of board that have both MicroPython and CircuitPython.
I am thinking about the following (1) Meowbit (2) Pyboard.
Do you see any other board?
gilded cradle
@half sedge, I believe the Feather STM32F405 Express does as well
tulip sleet
@half sedge Very early MIcroPython samd port done 6 months ago, missing many peripherals: https://github.com/micropython/micropython/tree/master/ports/samd/boards
Also overlap on nRF52840 boards: https://github.com/micropython/micropython/tree/master/ports/nrf/boards
pca10056, pca10059, particle xenon. (Feather52 is nRF52832: we no longer support). It would be easy (but not done) to port MicroPython to the various other Adafruit nRF52840 boards
idle owl
@lone axle After today's meeting, let's discuss getting you some hardware for testing.
timber mango
@tulip sleet Until now. I thought porting the full Micropython to other boards would be a good thing. Now, I am not so sure. The only feature I think would be useful that Circuitpython does not already have is interrupts.
pastel panther
@timber mango PyLint is a harsh but fair arbiter
timber mango
@pastel panther Indeed, she is. I will get better as I get tuned into what she wants. π
raven canopy
Negative scores are barely fair. π
pastel panther
π
idle owl
Negative scores are a badge of honor!
pastel panther
indeed
raven canopy
Collect em all!
tidal kiln
anyone ever got a NaN?
pastel panther
someone needs to make a t-shirt referencing PyLint scores
@tidal kiln no, that would be an achievement worthy of adulation
ivory yew
We definitely made pylint crash several times when I was working on the google client libraries.
timber mango
@pastel panther "PyLint is a harsh mistress"
idle owl
<@&356864093652516868> Weekly meeting is in just under an hour. Everyone is encouraged to attend! Please add your hug reports and status updates to the notes doc even if you'll be attending - it's super helpful. thanks! Looking forward to talking to you soon. https://docs.google.com/document/d/1FqFlX9sNJzdI8y41KqMu7kV0GPW8o7mzWxOoowAwOqA/edit?usp=sharing
ivory yew
feeding pylint generated code like
solar whale
nope
pastel panther
nor can I:
@timber mango yes, that was the reference, though personally I find the "harsh mistress" trope a bit tired
modern wing
With a Google hiccup, you can try the usual methods of treating it: gargling water, holding your breath, or even scaring it...
sneaks up on Google....then screams... AWS!
pastel panther
DEVALUATION!
modern wing
just checked -- nope didn't work.
solar whale
I want my money back -- oh wait, it's free....
ionic elk
solar whale
@idle owl broke google...
idle owl
Yeah I see that.
I can't even create a new doc.
We'll figure something out.
Came up for me.
modern wing
Back up for me too
timber mango
I have it also.
idle owl
@ionic elk @pastel panther @solar whale Appears to be back up.
lone axle
@idle owl sounds good, I'll be around.
errant grail
Text only today.
modern wing
Lurking today for myself.
onyx hinge
hope I don't have hiccoughs all through the meeting
solar whale
Run pylint -- that'll scare them away π
idle owl
I will have construction through the meeting, so... Might as well cough. π
onyx hinge
just voice chatting with @half sedge and I think it wouldn't hurt to mention the meeting time in UTC. except that we americans will get it wrong because UTC is practically metric
raven canopy
UTC? Is that band? π€ͺ
solar whale
A French Band
errant grail
Discord needs a local and universal time zone markup code.
half sedge
@onyx hinge Hopefully if every Monday it is the same time, then I can remember. But sure for newcomer it is a calculation to do. Maybe a PyBoard count down for next meeting (I believe there was a learn guide for that already).
errant grail
... or a Discord bot-based converter.
solar whale
I think UTC is a fine idea for all time references -- most of my computers are set to UTC.
onyx hinge
@half sedge the day of the week varies sometimes due to US holidays, but the time of day has been consistent. However, we switch summer/winter time at different dates than Europe does, which can also trip things up
half sedge
@onyx hinge EU is supposed to stop that summer/winter time within a year (or two?).
onyx hinge
Interesting -- many states in the US are discussing it too. Timekeeping is a fascinating topic, and timekeeping to serve people even more so
solar whale
reminds me of a favorite saying time flies like the wind .... fruit flies like bananas
idle owl
Google is your friend for figuring out time zones.
onyx hinge
hmm I had not considered that having a kmk (circuitpython) keyboard would disrupt my normal workflow for developing circuitpython code, since it takes over the CIRCUITPY mount point
idle owl
@onyx hinge Rename it?
onyx hinge
@idle owl hmmm that's a good idea, I think I know how to do that
half sedge
@onyx hinge Right now, it is undecided if we will go for summer time or winter time. The only time offset EU citizen need to know is GMT (UK/Portugal) and CET (everywhere else).
idle owl
@onyx hinge It maintains the name unless it gets corrupted.
tidal kiln
lurking
onyx hinge
yay that works, now it mounts as KEEB
thanks @idle owl I wouldn't have thought of that. just meekly accepted my fate
idle owl
You're welcome π
We're finishing up an internal meeting, we'll be there soon for the weekly.
errant grail
Leap year is also a time issue. My calendar gets messy every 100 and 400 years.
solar whale
Is that UTC soon or EST soon ?
serene warren
Lurking
gilded cradle
Please let us know in the chat if you're lurking. Thanks
modern wing
Happily lurking π
wraith tiger
I'm just lurking
old smelt
Lurking
indigo wedge
lurk-o-rama
onyx hinge
I noticed the RTC module in either the m4 or the nrf52, I forget which one, doesn't implement the divisible-by-100 rule of leap years
idle owl
Brb.
sterile bronze
lurking
timber mango
Lurking today
solar whale
Since 2000 was an exception -- no one cared....
timber mango
Text only, no mic.
solar whale
another 400 years to correct the algoritms
fierce girder
Is there audio now?
modern wing
Loud and clear @idle owl
solar whale
actually 2100 will be the next test case....
onyx hinge
@solar whale I don't think I'll be working on those bugs
errant grail
UTC leap seconds are concerning, too. They are added non-algorithmically.
indigo wedge
wow, the list doesn't fit on my screen
river quest
trying out this new emoji i just added 
obsidian compass
Dope
solar whale
looks like Lisa Simpsons hair....
indigo wedge
this is ing up to be a great meeting
timber mango
Forever!
tough flax
@onyx hinge - that works great!
river quest
also added 
Open Hardware Summit prototypes!
https://twitter.com/QwertyEmbedded/status/1220097577015603200
river quest
Speaking of... 2020 Open Hardware Summit #ohsummit20 topic on Adafruitβs Discord AND logos/art for the event with CircuitPython
https://2020.oshwa.org
https://discord.gg/8RVyqnk
https://www.dropbox.com/sh/8v0h6uariv04osr/AABHJ1wnEu0TOQ_mk0SivOtAa?dl=0
Mu version 1.0.3, 1.0.3 is a bug-fix release with no new features 
https://madewith.mu/mu/releases/2020/01/26/mu-1-0-3.html
Speaking of... Sponsor Nicholas Tollervey on GitHub Sponsors 
https://github.com/sponsors/ntoll
CircuitPython in HackSpace Magazine: Serpente 10 out of 10, and more... Three ways to blink an LED: Use CircuitPython to handle intermittent tasks by Ben Everard page 74 and 75.
https://www.tindie.com/products/arturo182/serpente-a-tiny-circuitpython-prototyping-board/
https://hackspace.raspberrypi.org/issues/27
turbid radish
Lurking
indigo wedge
π
river quest
Winners of the Take Flight with Feather contest!
The Open Book Feather - An open hardware reading device and more!
https://hackaday.com/2020/01/22/winners-of-the-take-flight-with-feather-contest/
https://youtu.be/iJqkwLt4Bik?t=54
CircuitPython + Open Book progress is here: https://twitter.com/josecastillo/status/1221622996353331202
slender iron
I like that discord has one account that is shared across servers
river quest
All a board! New boards on CircuitPython.org
Some of the latest boards to appear on circuitpython.org/downloads. The total is now 103+ boards!
https://circuitpython.org/board/aramcon_badge_2019/
https://circuitpython.org/board/kittenbot_meowbit/
ionic elk
These stats always make me nervous...
river quest
Speaking of... CircuitPython goes purrrfect with Meowbit
https://youtu.be/YFTbBri0rpc
half sedge
That make 3 Meow...
river quest
Grade 11 CS final project, Astroids clone, made with CircuitPython
https://ics3u-asteroids-davin.readthedocs.io/en/latest/
https://youtu.be/gHuDd-Jwamo
"Plug and Play Python (PnPP) on hardware IoT under 5 mins" ... "took less than 2 minutes and 40 lines of code!"
https://twitter.com/iAyanPahwa/status/1220810606862712833
obsidian compass
π
river quest
CLUE works with Bright Wearables
https://youtu.be/dx3qUkAB1ok
slender iron
@obsidian compass we don't have any google summer of code plans. We end up mentoring folks in our community as the opportunity arises rather than participating in a formal program.
river quest
All this and more in the DRAFT, please make an issue/PR for any updates, goes out 11am ET tues via adafruitdaily.com
https://github.com/adafruit/circuitpython-weekly-newsletter/blob/gh-pages/_drafts/2020-01-28-draft.md
obsidian compass
Thanks PT
turbid radish
If you are not subscribed to the newsletter, go to www.adafruitdaily.com and sign up - they go out Tuesday mornings.
onyx hinge
negative 2 issues not assigned a milestone?
timber mango
@idle owl I am interested in starting to look at PRs for libraries. Does that have a list of all libraries with PRs?
lone axle
@timber mango this page has a list of the currently open ones: https://circuitpython.org/contributing
timber mango
@lone axle Thank you.
graceful heart
late, but lurking/text only
timber mango
Low audio.
inland tusk
Lurking for hug reports
lone axle
π
obsidian compass
I have a hug report
timber mango
Is audio gone? I do not hear anything.
slender iron
@timber mango I have audio
try restarting
@obsidian compass would you like to be in circuitpythonista's? that way you'll get pinged with meeting info
obsidian compass
@slender iron sure
slender iron
cool, done!
timber mango
Restarting worked.
slender iron
@onyx hinge mute please
onyx hinge
@obsidian compass I apologize for skipping over you, I was going off the notes doc and should have been looking at the list in the voice channel as well
slender iron
unless you are running status updates π
obsidian compass
@onyx hinge no worries, I shall add to notes as well π
gilded cradle
Thanks @obsidian compass, I missed typing one of your hug reports
tough flax
Did I miss my turn? Had a day-job issue
slender iron
your hug was read off
tough flax
Great - thank you!
obsidian compass
@gilded cradle added in the doc π
slender iron
π
onyx hinge
dream big, and nibble some things off the infinite list
gilded cradle
Dark Magic π
modern wing
I can't tell you how many of those "would be nice" items on the infinite to-do list have had the most impact. I wonder how I've gotten by without them.
slender iron
@tough flax the gaming folks are in pittsburgh right? that is where pycon is the next two years
tough flax
@slender iron - AbleGamers (who we're releasing w/this week) has a presence in Pittsburgh
Yes, we should plan to play there
(Technically they're in West Virginia, but Steve Spohn is in Pitt)
slender iron
ah neat. super happy you are doing stuff together
tough flax
Very cool stuff releasing this week - save room on the blog π
gilded cradle
That's pretty cool @obsidian compass
inland tusk
@slender iron I would like to talk to you about the keyboad.
slender iron
we always have room for it π
@inland tusk nice! sounds like you are getting ahead of me on the project
obsidian compass
Thanks @gilded cradle , it's really cool, fun and easy π
gilded cradle
Excellent to hear
obsidian compass
Need to rush, thanks everyone π
timber mango
Circuitpython reminds me of a steam roller, the way it has taken off, along with the Feather ecosystem. Awesome stuff, Adafruit!!
fierce girder
Re: BeagleBone Green https://github.com/adafruit/Adafruit_Python_PlatformDetect/pull/60
slender iron
yay! both good things
raven canopy
Frustratingly Fun. I'm in the same space, @solar whale!
errant grail
My Funometer and Frustometer generally track each other.
wraith tiger
The frustrations make finally getting it working feel like more of an accomplishment.
timber mango
@onyx hinge audio is very low.
fierce girder
@gilded cradle I closed this PR as my BBG conflicts with it
https://github.com/adafruit/Adafruit_Blinka/pull/201
I need to open a new one once I refactor how to add support for PWM on the BeagleBone Black
I need to modfiy src/adafruit_blinka/microcontroller/am335x/sysfs_pwmout.py so that it uses these deterministic /dev/pwm entries in order for there the correct mapping between the P8/P9 pin and the corresponding peripheral.
timber mango
@wraith tiger The intensity of the buzz I feel when getting a really difficult problem solved is much higher.
onyx hinge
@timber mango I'll work on it after the meeting
half sedge
It would be great to be able to compile from .py to .mpy on CircuitPython M4 hardware. Even if it does not solve your problem.
onyx hinge
I have to step away for a few minutes
tough flax
Gotta drop - was staying for that @onyx hinge note - but I have a stupid health insurance meeting for my employees (the horror!)
median ledge
Out of curiosity, why is circuit python voice flooded?
fierce girder
@gilded cradle i've created this if you can assign to me https://github.com/adafruit/Adafruit_Blinka/issues/240
slender iron
@median ledge we have a weekly meeting every monday
median ledge
ah
turbid radish
We have many outside guide authors. We have not contracted them for extended support so Adafruit does the best they can
onyx hinge
eek that's not a happy cat
median ledge
neat
when I used to work at boeing, the machining floor also had monday meetings to give an outline for the week, works pretty well
serene warren
We had "Tier X Meetings" Tier 1 was something that could be done by the worker bees, Tier 2 needed a management decision, Tier 3 was Deprtment level, and so on.
median ledge
cool
onyx hinge
time to admit it's a 90 minute meeting
modern wing
Thanks for another awesome meeting, catch y'all next time!
errant grail
Thanks everyone!
serene warren
Great meeting folks. I never did find the Notes doc though
half sedge
Notes should be sorted by time now. I have one timestamp missing (search FIXME). @idle owl
tulip sleet
@serene warren Click the "pushpin" icon at the top for a pinned messsage with a link to the doc
gilded cradle
π
raven canopy
π
wraith tiger
π
graceful heart
π
serene warren
Thanks @tulip sleet , I'm always leary of pushing buttons
idle owl
@half sedge Thanks!
onyx hinge
π
gilded cradle
Thanks
solar whale
π
tulip sleet
@slender iron I hav a q re remote service characteristics when you are back
I brought DeviceInfoService up to date (mostly adding a service kwd arg, but when it binds to the remote service it attempts to match all its characteristics with the remote, and some are missing, and it throws an exception
trying to decide what the right thing to do here is
solar whale
@slender iron @tulip sleet perhaps we could consider a small C interrupt handler to capture RFM9x/69 packets - the maximum packet size is ~250 bytes --- I'll think about this more and make a specific request/proposal if it seems like a reasonable way to go. The code to handle them would not be very much since the "ACK" packet is a fixed response. Something to consider....
slender iron
@solar whale I'm open to that.
tulip sleet
@solar whale I wonder if this could be generalized into a "do this SPI transaction on interrupt"? Might be useful for other devices as well. If the transaction doesn't need to change based on something, yes, if it's dependent in some way, then no.
slender iron
@tulip sleet back now
tulip sleet
so I think the DIS on the HRM does not have all the characteristics
solar whale
Thanks -- I'll think about it and get back to you.
tulip sleet
def _ensure_bound(self, service, initial_value=None):
"""Binds the characteristic to the local Service or remote Characteristic object given."""
if self.field_name in service.bleio_characteristics:
return
if service.remote:
for characteristic in service.bleio_service.characteristics:
if characteristic.uuid == self.uuid.bleio_uuid:
bleio_characteristic = characteristic
break
else:
raise AttributeError("Characteristic not available on remote service")
else:
bleio_characteristic = self.__bind_locally(service, initial_value)
service.bleio_characteristics[self.field_name] = bleio_characteristic
@tulip sleet shouldn't that cause the attribute error when the service's characteristic is first accessed by user code?
tulip sleet
oh, wait, you're right. I was looking at the wrong line in my user code. I thought it was the conn[SomeService] line, but it's not. I was working on this during the 1pm meeting.
ok, nvr mind
sri
idle owl
@half sedge The missing time code was earlier in alphabetical order - that person's hug report was in two places. One time coded, one not.
idle owl
@lone axle and anyone else interested in contributing on PRs: please make a comment on a PR if you are going to be testing it. Having multiple people testing the same PR is great, but it's also great if there are other PRs that someone could test that we get more PRs reviewed as well. So make a note when you begin testing, but feel free to test the same PR as someone else as well if that's the one you're interested in/able to test.
lone axle
Will do.
onyx hinge
hmmmm if there are both .py and .mpy, it seems .py is preferred
(at the same directory level)
solar whale
Yes
idle owl
@raven canopy Apparently Adabot is including archived, read-only repos in the list of libs that have commits since last release. I think we've only very recently had any read-only CircuitPython libraries, so this is a new situation.
I'll file an issue but I assume we need to add something to have it ignore archived repos.
raven canopy
raven canopy
was supposed to exclude archived repos... but, i'll verify. π
oh. wait. circuitpython.org's adabot submodule does still need to be updated. so if that's where you're seeing it, that's why.
PR imminent
idle owl
Thanks!
idle owl
@raven canopy Missed that your comment was half question there. Yes that's where I was seeing it.
raven canopy
hehe. all good.
lone axle
@idle owl are you around still? I'm about to work on editing the file for that PR we were discussing previously.
idle owl
@lone axle For a bit yes. Dinner will happen eventually.
lone axle
@idle owl Okay, I think I'm all set now. I went to the file, clicked edit. I've made my changes and now I have a box at the bottom to type a commit message and button labeled "Commit Changes"
idle owl
There should be an option to "create a branch and start a PR" or some such
lone axle
Okay, yep I see that one
idle owl
Do that.
idle owl
Should create a "patch" branch and try to merge it to the PR
lone axle
I pushed the green button at the bottom and thought for a second then took me to github 404 page =x
lone axle
Tried once more just incase it was a weird github fluke but same results
idle owl
So you clicked "Propose file change" ?
and it 404'd?
Hey!
It did for me too.
that's new and fun.
lone axle
yep.... uh oh
idle owl
Hmph.
I guess make a comment on the PR that you have the fixes for it but GitHub is acting up.
I tried on another repo, and same thing.
lone axle
Caught it in the inspector the second time around I can see that the POST to change the file is getting back the 404 response. But not much else in the way of helpful info unfortunately.
Ok, will do.
idle owl
Actually, try one more thing for me:
Add your changes and commit directly, since you said you had access.
Don't create a PR
it will run the CI testing on your changes if you commit directly to a PR
so it's not ideal, but it doesn't mess with anything
lone axle
Ok, leave the radio button on " Commit directly to the master branch." then for this one right?
idle owl
Yes. Click the button.
We'll see if the GitHub GUI is entirely broken, or if it's only the PR creation
lone axle
done. It seems to have worked that time.
idle owl
Alright! It worked.
So now we let the CI run and see where we're at.
once it passes, I'll tag Scott again because he reviewed it last.
Thank you!
lone axle
Yep your welcome, thanks for the help working through it.
idle owl
Alright, there's some pylint errors. Which is how it goes π
probably need to add the attributes to __init__ and define them as something, possibly None. But you'll want to test that it doesn't break anything.
Do you see how to get to the Pylint errors?
Presumably yes since you've fixed some already.
lone axle
Yep, found it
idle owl
Under normal circumstances I would have you setup Pylint locally if you haven't already, but now we're working in the GitHub GUI unfortunately, so it means letting the CI do its thing.
Make your changes the same way and commit them to master the same way.
lone axle
Okay, I can clean those up
lone axle
Hmm, although two of them are referencing self.x and self.y. I didn't create those, they would have come from the Group superclass I think?
idle owl
hmm.
They're called x-self and y-self in the property
and in the setter, x and y . Is that right?
lone axle
_anchor_point I did create, and I do have it defined in init. Perhaps I'm not understanding what the pylint error means.
idle owl
No, I think you're understanding it properly.
lone axle
ah, yes you are right x-self
idle owl
But I'm wondering if it's not quite right.
lone axle
Actually pylint shows the errors on lines 199 and 200, which don't use the x-self and y-self
idle owl
Right.
lone axle
Oh, I understand the issue with _anchor_point actually. When I copied over my code I put in all 4 methods but not that line from the init function
idle owl
π It happens
lone axle
The x, and y ones I'm not sure what to do about though
idle owl
Does it work as is?
lone axle
I can try setting self.x and self.y to 0 in the init but I don't quite understand all of the displayio / group system, not sure if thati s going to cause trouble
As far as I can tell it is working (once I put the line back in the init function)
idle owl
Alright, add the line back to init. Beyond that, this is a case where pinging someone who is more familiar with displayio might make sense. I would suggest pinging Scott or Carter on that PR to ask for feedback.
@tannewt or @caternuson on GitHub.
To find out whether it's ok to simply initialise x and y, or if something else should be happening there.
lone axle
Okay I've got that line added back in. It is now the same as my local copy is working for me. I'll go ahead and add a comment with the test code that I used as well.
idle owl
Sounds great!
idle owl
@lone axle Once you're done with that, let me know and we'll go over what other PRs you want to test that you need hardware for.
Or it might have to wait. Dinner prep is starting.
@lone axle I may be back later, or we'll go over it tomorrow. I'll ping you later if I make it back. Otherwise, I'll ping you tomorrow and let you know when I'm around.
lone axle
@idle owl ok sounds good.
timber mango
I am still having a HUGE problem between Circuitpython (5.0.0 beta.4) and the Arduino 1.8.10 IDE. I do not know whether this is a problem with the IDE or Circuitpython or something else.
This message appears after the device is reset after programming.
slender iron
@timber mango what shows that warning?
the BOOT drive disappears after the device is programmed and is reset
timber mango
I think it is something from Circuitpython.
slender iron
circuitpython can't show a warning on your desktop
timber mango
It is like it is catching the reset and wanting to go to bootloader mode.
I do not know where the message is coming from then.
raven canopy
so you're flashing an Arduino program to a board, after it previously had CircuitPython firmware flashed onto it?
slender iron
sometimes computers complain about the boot drive disappearing. it's not a problem because the bootloader knows when it is safe to reset
timber mango
@raven canopy Yes, a Feather M0 Express.
I can only get very small Arduino programs to run.
It does seem to be a problem in my case though. I have tried after a fresh reboot of my computer.
raven canopy
have you tried programming it in the normal mode (not "FEATHERBOOT")?
slender iron
the ide can trigger the bootloader
timber mango
It seems to be triggering the UF2 bootloader.
I do not know though.
What would trigger Linux to show that FEATHERBOOT message?
lone axle
@timber mango when you plug in another drive like a flash drive or external hard drive, does it automatically open up a window showing you the files?
timber mango
It does that when I plug in a Circuitpython device or any device that can be mass storage.
lone axle
I think that is the culprit. The board is going into bootloader mode for a few seconds while the IDE writes your program to it, then switching back to non bootloader.
Your OS sees the bootloader drive exist for a second and tries to open up that window for you. But by the time it's ready to show you the files the drive is gone already because it has switched back to normal mode
timber mango
Possibly. The delay is enough to upset running the Arduino program.
lone axle
I would probably try to disable the auto launching window of plugged in drives all together or at least on the bootloader drives if you can choose them by name or something
timber mango
If I could do that, and I do not think I can except maybe globally, I would probably have to manually mount my Circuitpython drives. If that could work it would potentially mess other things up too. Is this what you are referring to (disable automount for everything)?
lone axle
Ideally not disabling the automount, just the UI window that pops up trying to show you the files
slender iron
timber mango
I am pretty sure I do not have that kind of control.
lone axle
What kind of Linux / UI are you using?
timber mango
Raspbian Buster on a Raspberry Pi 4.
Everything used to work, but not anymore since I plugged a Circuitpython device in at the same time as a device I use Arduino with.
slender iron
@timber mango I don't think it has to do with circuitpython. Its the new UF2 bootloader which shows up as a storage device. The previous bootloader was serial only
lone axle
I think this thread might be talking about a setting that should be relavent: https://www.raspberrypi.org/forums/viewtopic.php?t=159739
a few posts down someone talks about disabling "Show available options for removable media when they are inserted "
it's a few years old so the name could have changed or where to find it perhaps. But I bet there is setting like that around somewhere still. If you can find it I think it would stop the pop-up about failing to find the drive from showing. However as you noted it would affect all USB storage, so you'd always have to explicitly open the filemanager when you want to look at files on a USB device.
meager fog
@slender iron i also noticed on sunday they deployed a new interface
timber mango
@lone axle I am two versions later in my Raspbian and I do not see any of those options in my GUI. I do not use the standard Raspbian GUI.
I am going to have to google around and see what program might do that for me,
bright aspen
On macOS Mojave and Feather M4 Express with CircuitPython (something between beta 3 and 4) when I download a file the main gets loaded twice.
half sedge
Hi, I made this to promote my talk at Fosdem'20: https://twitter.com/DavidGlaude/status/1221954266887270402?s=19
Could it be used to put in the newsletter of tomorrow?
timber mango
I was able to turn off automount for all disks, but that did not have any effect on the FEATHERBOOT message. π« I need to give up on this for tonight as my funometer is not high enough.
That FEATHERBOOT message must be classified as an error.
craggy galleon
Does CircuitPython have SBUS support?
craggy galleon
I think a chip would be required.
craggy galleon
I don't know too much about it except that in RC control it sends 8 PWM signals on one connector.
It seems that FrSky X8R and XM+ are a popular choice
simple pulsar
What is Sbus? I suspect it's not the one I'm thinking of
river quest
CircuitPython - The easiest way to program microcontrollers
https://circuitpython.org/
https://youtu.be/40gZBLdDSZA
obsidian compass
Wow, new contributors coming in! Some of em are attendees from CP day we celebrated in Delhi last year π
Great to see them creating issues
tulip sleet
@timber mango That message is really nothing to worry about, though it's a nuisance. I'll try to reproduce it. For what it's worth, I don't see that on Ubuntu 18.04 running cinnamon desktop. It may have to do with the file manager that you are using on Raspbian. Are you using the standard Raspbiran desktop or something else?
orchid basinBOT
main meteor
@simple pulsar I suspect you and I are thinking of the same (Sun) SBus (which I designed this card for in the past), but I think you're right that it's not the one under discussion here.
obsidian compass
RC receivers these days outputs SBUS/IBus/CRSF kind of digital serial signals
PWM and CPPM are kinda outdated now IG
stuck elbow
"SBUS is an inverted UART communication signal."
looks like one transistor or not gate, and you are good?
obsidian compass
Yeah, More or less,
100000 baud rate, 25 bytes message
It was developed by Futuba IIRC which is an amazing RC transmitter manufacturer company, itβs quite reliable
stuck elbow
by the way, are you Chumanista?
obsidian compass
No Iβm not
stuck elbow
I see, sorry, that's a guy who has a youtube channel called "code and solder"
obsidian compass
Aah, I didnβt know that. I go by my name Ayan Pahwa | codeNsolder on YouTube and across web.
Also @slender iron has developed some drone flight controllers in the past so he should be knowing a lot about Rc stuff π
orchid basinBOT
bright aspen
@onyx hinge I've been working on a pure tone generator project, but last night switched back over to looking at SDcard. I switched to another Class 10 card and the time for a random read took 2-3 ms longer. Do you have a criteria for SD cards for audio?
onyx hinge
@bright aspen no, I've just been using whatever adafruit stocks, https://www.adafruit.com/product/1294
afk for a thrilling visit to my dentist
bright aspen
@onyx hinge Ah, I might have one somewhere. I switched over to this one: http://adafru.it/2693
It says SanDisk Ultra 10 GB MicroSDHC I Class 10 A1. A1 is supposed to be good, right? And the I, that's good, too?
@obsidian compass Though it might not help with your current task, I think adding pin polarity to UART is a reasonable enhancement suggestion, should you want to make it. (I have no idea what implementation problems might be, though.)
prime flower
wow, the new photos/landing page on the website came out great
obsidian compass
@ivory yew congratulations on talk selection at PyCon. Iβm really excited for that talk
Hope I can make it to the conf π
idle owl
@tidal kiln We ran into issues trying to add new files to that PR, so if you approve it, I'll merge it and we can start a new PR for the example. @lone axle
@tidal kiln I realised after commenting that you had an active change request still, so I didn't want to merge it out from under you without approval from you first.
tidal kiln
@idle owl is the PR urgent? if not, will suggest one more minor thing.
idle owl
It's not urgent, but bear in mind that FoamyGuy is editing everything through the GitHub GUI because the original author is not currently involved. So if it is a change to anything but the currently active file, it will require a new PR.
Otherwise, suggest away π
tidal kiln
@idle owl meh. just made it a comment suggestion. approved.
idle owl
Fair enough! Thanks π
stuck elbow
would it make sense to enable this for the circuitpython.org website repo? https://github.com/marketplace/imgbot
ivory yew
@obsidian compass thank you!
idle owl
@lone axle I will be around in about 15 minutes until a 1:15 meeting that will likely not be more than 30 minutes, and then I'm pretty free for the rest of the day. If you're around today, we should chat.
@obsidian compass If you end up making it out to PyCon, make sure to let us know so we can meet up.
obsidian compass
@idle owl yes, thatβs the plan π
lone axle
@idle owl Works for me, ping me when you're ready.
slender iron
@obsidian compass it has been a while since I've done RC stuff π
obsidian compass
@slender iron haha yeah, Iβve used your fc stack on one of my freestyle build in the past π
slender iron
oh man. I hope it worked ok
obsidian compass
It does for a while and then I crashed into a river, was flying real low :p
solar whale
@gilded cradle FYI -- just got blinka up and running on an old beaglebone black that had been gathering dust -- just hooked up my first I2C device -- working well so far. Very nice!
gilded cradle
Thanks @solar whale, @fierce girder has been doing most of the work maintaining that board.
ionic elk
@slender iron quick question - I've been looking at the guide for the hallowing m0, but there doesn't seem to be any example code for actually using the screen. What's the intended access method, once you've defined it in board.c?
stuck elbow
@ionic elk what do you mean? Usually you have it in board.DISPLAY and can use it with displayio
@ionic elk here is a snake game for the hallowing, maybe that will work as an example? https://github.com/deshipu/hallowing-snake
slender iron
@ionic elk as @stuck elbow said, the Display object is available via board.DISPLAY
tidal kiln
meager fog
@indigo wedge hihi ok good news - the PCBs i got dont have a 3.3-GND short
indigo wedge
@meager fog yeah my pcb fab admitted it's on them and they're remaking them for free but they won't ship until friday
meager fog
@indigo wedge you wouldn't by any chance have a PDF/PNG stuffsheet would u
ive never used kicad so it would be an adventure 4 me
indigo wedge
Stuffsheet?
Like assembly drawing?
I can send you PnP, bom and assembly drawings when I get home in a bit
onyx hinge
I'm trying to get up the learning curve on kicad too. found out the engineering college is having a tutorial session on it tomorrow, but I can't make it π
idle owl
@lone axle If you're around, I'm free.
indigo wedge
@meager fog I think master is untested right now on the feathers so might require some tinkering to get it to work, same for the uf2 bootloader, I'll look at it on Monday when I get the pcbs and assemble the first one
lone axle
@idle owl I am here
idle owl
Hello!
Ok.
Let's talk PRs. What PRs are you interested in reviewing but don't have hardware for?
lone axle
I think I could handle testing these if I had the hardware. I've added a brief note about my understanding of what would need to be done
https://github.com/adafruit/Adafruit_CircuitPython_BME680/pull/18
Gas sensor. Test the sensor on CircuitPython with and without the changes from this PR. Test sensor with Arduino to compare results.
https://github.com/adafruit/Adafruit_CircuitPython_VL53L0X/pull/12
ToF Distance sensor, adding support for multiple sensors via I2C. Changes need test with multiple sensors.
idle owl
Alright excellent.
lone axle
As long as my understanding is correct then I'm pretty sure I can handle the testing for them
idle owl
Ok, first one has been merged it looks like.
Second one, your assessment seems accurate.
lone axle
Oh, yep looks like some more work has been done on it since I went through them and made my list
idle owl
Ok, I'll explain how this works via DM.
ionic elk
@slender iron one other quick question - what should I pass in instead of NO_BRIGHTNESS_COMMAND? All the example boards seem to have auto brightness set but I just want the meowbit to default to 1 for now.
slender iron
Auto brightness defaults to 1 now. Brightness command is only for oleds
idle owl
@main meteor Ping me when you're available for a chat.
ionic elk
ok
ionic elk
My jlink is returning 0xfffffffe in ?? () anybody run into that before?
I can't get it to do anything else, even on reset
onyx hinge
is displaying text items just really memory intensive? I see that each character gets its own TileGrid, which is maybe this 84 byte structure whose allocation keeps failing in my program?
ionic elk
I think text is generated in RAM so that would make sense
onyx hinge
it's bumming me out right now.
onyx hinge
π¦ π¦ any spares?
ionic elk
nope
onyx hinge
free memory goes from 55264 to 12224 displaying this text-based menu of a few hundred characters. eek.
"""Display a menu and allow a choice from it"""
gc.collect()
print("***", gc.mem_free())
board.DISPLAY.auto_refresh = True
scroll_idx = sel_idx
glyph_width, glyph_height = text_font.get_bounding_box()[:2]
num_rows = min(len(seq), board.DISPLAY.height // glyph_height)
max_glyphs = board.DISPLAY.width // glyph_width
labels = [adafruit_display_text.label.Label(text_font, max_glyphs=max_glyphs)
for i in range(num_rows)]
cursor = adafruit_display_text.label.Label(text_font, max_glyphs=1, color=0xddddff)
base_y = glyph_height//2-1
scene = displayio.Group(max_size=len(labels) + 1)
print("-> ", gc.mem_free())
for i, label in enumerate(labels):
label.x = round(glyph_width * 1.5)
label.y = base_y + glyph_height * i
label.text = seq[i][:max_glyphs]
scene.append(label)
print("-->", gc.mem_free())
always returns
#0 0xfffffffe in ?? ()
#1 <signal handler called>
can't get it out of that, no matter how many flash erases and resets I try
onyx hinge
you de-powered them both, and restarted gdb and jlinkgdbserver programs?
onyx hinge
20 labels, total 313 characters, uses 43kB RAM
ionic elk
does it store each of them as a tiny duplicate bitmap or something?
onyx hinge
I don't know the specifics of how adafruit_display_text.label.Label works
however, I THINK that each character from the font which is loaded is a separate Bitmap, each Label is a Group of some size, and each item in that Group is a TileGrid which shows one of the bitmaps from the font. It ends up being a lot of fairly bulky items
@ionic elk they'll surely get you a fresh meowbit, they're in stock
granted that may leave you a bit stuck for a few days, which is why I've been recommended to always get 2 of something I'm working on
slender iron
@onyx hinge you could try using my heap analysis script
onyx hinge
@slender iron good idea
slender iron
its fun π
that seems kinda high to me since the bitmaps for characters should be deduplicated
onyx hinge
as a starting point, is this a shocking amount of memory for something "simple"? I'm not just miscalibrated, am I?
slender iron
but then again displayio is designed to minimize how much is transmitted to the display
onyx hinge
this started causing regular MemoryErrors in jeplayer after that "other" buffer got doubled from ~2.3k to ~4.6k so apparently I've been close to an edge
onyx hinge
I don't know. One failed memory allocation said 84 bytes.
slender iron
(gc allocs in 16 byte chunks)
the areas could be pretty big
you could use a debugger to stop when an allocation fails
onyx hinge
can you send me that link about how to use the script? I'm not finding it right away.
analyze_heap_dump.py ?
onyx hinge
To dump ram do this in GDB: dump binary memory ram.bin &_srelocate &_estack
found this little note
slender iron
ah yup
onyx hinge
that's all literal text to type?
slender iron
ya
onyx hinge
back to the debug-cave where I keep my segger
slender iron
π
that will output the ram memory range into a .bin file
it is definitely hacked together
but it'd be good to get you ramped up on it
onyx hinge
learning is good
slender iron
oh, and I think I was counting words so 128 bytes per tilegrid sounds about right
if the font is monospace then you could use a single tilegrid
onyx hinge
it seems like (A) a GlyphCache font has one Bitmap per loaded glyph, and (B) a label has one TileGrid per displayed character
which (B) adds up to a lot
slender iron
ya, I believe it
onyx hinge
but because the glyphs are spread over multiple bitmaps I don't think an approach with a single tilegrid will pan out
slender iron
are they different fonts?
onyx hinge
no, but a bdf font creates a bitmap each time it loads a glyph .. I think ? https://github.com/adafruit/Adafruit_CircuitPython_Bitmap_Font/blob/master/adafruit_bitmap_font/bdf.py#L150
slender iron
ah, ya it does. we could have it load it into one if its monospace
thats what the internal font does
onyx hinge
if the glyphs could all be in one big long bitmap, but then you'd have to size the bitmap at the outset
or have .. disjoint bitmaps?
or let each TileGrid cell refer to a different bitmap?
onyx hinge
"can load up to 256 8x8 glyphs" e.g.
slender iron
right
onyx hinge
I can't directly write to the terminalio's grid, can i?
I'm shifting between "what can I do for jeplayer" and "what can I do for everyone" mode here
slender iron
no but you can create your own terminal
which uses one tilegrid under the hood
terminal just handles the input stream
onyx hinge
looks like there's no "go up and start typing on the top row again" type code
onyx hinge
I can reset cursor X via an escape code but not cursor Y
onyx hinge
standalone test for pygamer or probably any board: https://gist.github.com/953c4fceae40614effe693fcb13c9dd2 typical output: https://gist.github.com/d25e003a46b0ffdf9509c0b1a4765ef4
displaying a 32 character string takes about 3462 bytes
@slender iron I'm not sure that adafruit_display_text can do better without changes in the core (I don't know what those changes are though). should I file any issues, and if so, where?
slender iron
Β―_(γ)_/Β―
what do you expect it to do?
it's a constrained system. at some point you'll hit the edge of what it can do
onyx hinge
"Shelly's application needs a way to display hundreds of characters on-screen, with RAM usage per additional on-screen character on the order of 1-4 bytes, not 100 bytes"
ionic elk
@slender iron how do you tend to trace drops to the reset handler that don't pass through hardfault? trying to track a crash related to initializing fourwire in board.c
I don't have a MTB set up and that seems kind of like overkill. Is there some way I could just step until Reset and dump the results into a file?
@onyx hinge @slender iron is there some kind of flattening operation that could occur, where the raw aggregate of several layered bitmaps is stored and the components are discarded?
slender iron
I've gotta go meet a friend. I'll be back later.
ionic elk
@slender iron when you get back: the reason that we can't do a board.c initialization of the screen is because the internal_spi bus doesn't actually exist at that point. So unless we want to introduce a way for board.c to be executed after the initialization of the SPI used in the filesystem, the meowbit will need the screen manually defined. I think.
half sedge
Hi, have in mind that someone made a LED animation library. I was thinking it was fancyLED, but it is not. Maybe the library I am looking for does not exist. My souvenir is that it was demonstrated on a xMas tree... Was it @idle owl or someone else? Did I dream or misunderstood? Do I confuse with _pixelbuf?
It would be something that "work" in the background and help doing asynchronous thing because you don't have to handle the animation once launched.
If anybody knows, please tag me.
onyx hinge
@half sedge you may be thinking of https://learn.adafruit.com/circuit-playground-bluefruit-neopixel-animation-and-color-remote-control
@slender iron I found a way that my code can use terminalio and memory usage is much lower as we expected.
half sedge
Yes, there is animation stuff. And I guess it is there to simplify handeling both BLE stuff and LED stuff in parallel.
onyx hinge
(a bunch of 1-line terminalios have much lighter memory usage and the fact that I can't work across multiple lines is just fine in that case)
idle owl
@half sedge That was my xmas tree. I haven't posted it anywhere, it used a combination of an unpublished version of the LED Animation Library and Pixelbuf. You didn't dream it, but it's not posted anywhere. Pixelbuf is now in flux to fix it, and there's no point in trying to get the LED Animation lib updated until that is complete.
half sedge
@idle owl I have a good memory, but I distruss my understanding of oral english, so it was not clear exactly. Is that animation stuff writen in C because that is the only way to do parallel processing?
idle owl
@half sedge The pixelbuf portion of it is written in C. I do not understand it well enough to explain it in detail, but essentially C is faster, so it moves all of the LED processing to C. So the animations are significantly faster as they are being processed in C.
onyx hinge
@ionic elk terminalio requires that the font be a "BuiltinFont". A BuiltinFont's glyphs all reside in a single Bitmap and so you can display any of the characters in a single TileGrid. So, basically, a bunch of intermediate Python objects required by the Label class aren't needed, but you can't use a font loaded from disk
idle owl
The animations themselves are written in Python. But the LED processing behind the scenes is written in C.
ionic elk
@onyx hinge that sounds like a problem that could be solved, if there's need for it
half sedge
Well, I really think this should be published and documented. It sound great. Of course with the breaking that _pixelbuf did, it may have to wait.
I can get that moving things in memory can go faster in C.
idle owl
That is the plan, but until pixelbuf is completed, it must wait.
half sedge
I have some colorfull reflexion. I found that FancyLED that can produce palette, and in the MLX thermo camery example I did, I build palette for displayio using a totally different way. So I was wondering if I needed to replace thing.
Then in many example, there is that function that produce "rainbow" (or transition from red to green to blue in circle) that is used for anymation.
There is that wheel function that is used and copy pasted in various example.
idle owl
As far as I know, it's built into pixelbuf. So you can call it without the function. I may be misremembering that as something someone wanted and it didn't get added.
half sedge
Maybe that should be put somewhere (the wheel function) and reused. Or use FancyLED to do similar effect.
lone sandal
idle owl
The plan was to build it into pixelbuf so you can call pixelbuf.color_wheel or some such and it works.
half sedge
I'll look at the code one day. I am supposed to know C and be able to write code in C. At least I did that in the past. π
Great plan.
Having the Python version is still good for education, and then say that the library is doing that in C for whatever reason.
solar whale
@odd flax Interesting news https://blog.particle.io/2020/01/28/mesh-deprecation/. See the end where they recommend installing CIrcuitPython
idle owl
The Python version is in many guides for posterity.
half sedge
I guess many CircuitPython user have Python as their first language.
idle owl
CircuitPython was my first language. π
solar whale
Mine was FORTRAN π
half sedge
It is good to learn another (or multiple other) languages. This give a perspective as a single language might limit your vision.
onyx hinge
I have a feeling I'm still un-learning habits picked up with 80s home computer BASIC, which is where I started.
half sedge
Python is my last language, but I have not learned it a Pythonic way, so I am affraid I writte Python code with C style (not using iterator, using index in table, ...).
I would love to find a way to forget the easy way (for me) to write code and learn how it is supposed to be written in Python.
idle owl
Habits are difficult to break π
half sedge
Except that they removed GOTO and that help avoiding it.
Ok, I'll stop playing with LED and figuring out about those various lib. This does not have an direct use for my presentation, so it can wait for after that.
Thank you @idle owl , and by the way, nice Xmas tree. π
idle owl
You're welcome. And thank you!
raven canopy
Sad news about Particle Mesh. But that was a very well constructed "why", and the store credit is about as much honest-dealing you can ask for from a company.
manic glacierBOT
solar whale
@raven canopy and the xenon works very well with CP.
umbral dagger
Are there anmy benchmarks comparing CP on the Feather M4 Express and Feather nRF52840 Express?
timber mango
@umbral dagger That would be a great item on a comprehensive comparison chart of Feather/Express boards.
slender iron
@onyx hinge I don't think you want terminalios. You just want tilegrids that are more than one glyph (which is what terminals use). It'd be cool to have the bdf loader create one bitmap or have tilegrid store a bunch of bitmap pointers. then you could load monospace fonts off of the disk
old smelt
Did the CircuitPython newsletter go out today? I normally get it shortly after 10:00am CST, but not today.
slender iron
@old smelt nope, they debugged it and are waiting to see if it goes out tomorrow
onyx hinge
@slender iron I agree with the sentiment, what I implemented is not a good technique that we would want people to use generally. For now, I may have to stick with terminalio for now as a matter of pragmatism. Would you like me to file an issue with the long-term milestone about the problem, while the discussion is fresh in our minds?
slender iron
I don't know what the milestone would be. TileGrid takes memory because it allows you to do a lot of things like per-glyph positioning, rotation, and indexing into the bitmap.
It should be straightforward to use a single TileGrid instead of terminalio. Is the confusing part mapping a character to an index into the bitmap?
onyx hinge
My original used Adafruit_CircuitPython_Bitmap_Font + Adafruit_CircuitPython_Display_Text, and now I am using the in-memory font and terminalio.Terminal. It's true, I might be able to use in-memory font and my own TileGrid class.
The issue would say, "we need to provide a way that Adafruit_CircuitPython_Bitmap_Font + Adafruit_CircuitPython_Display_Text can get memory usage almost as low as terminalio (plus the actual storage of the font)", though I don't think we know exactly what hte best course is
in-flash font
slender iron
I don't think that is concrete enough. I also don't think it's easy
terminalio saves memory by assuming that the font glyphs are a uniform size. The Bitmap Font and Label classes don't make that assumption
lone axle
@idle owl https://gist.github.com/FoamyGuy/0455545c888aaa408cdcbc54fd33774e I took a look through both lists and was able to find a few additional ones. Ping me tomorrow when you are around.
onyx hinge
If I'm not proposing a concrete solution you'd rather I didn't file an issue at this time.
slender iron
I think there are some options
- add the ability in bitmap font to produce a single bitmap for a set of characters
this will allow for tilegrid to show multiple characters from a loaded font
- In core, add a stripped down version of TileGrid that only works with a single bitmap. (Or split the implementation so that unneeded stuff is in a second allocation.) This will only get you down to 32b or more per object though
onyx hinge
I do now see what you said, that Display_Text supports non-monospace fonts
slender iron
and is orders of magnitude harder with layout
onyx hinge
I think that the fact that a font has a get_bounding_box() method made me think they were monospace only
slender iron
nope π
I really want to get joeycastillo to create awesome text layout for us
would love to do other directions
onyx hinge
There have got to at least be ideas they can share with us
if a font is a BuiltinFont then is it monospaced? Or is that not true either?
umbral dagger
Re: benchmarks. Here is is for PyStone:
Adafruit CircuitPython 5.0.0-beta.4 on 2020-01-22; Adafruit Feather M4 Express with samd51j19
Pystone(1.2) time for 500 passes = 0.34375ms
This machine benchmarks at 1.45454e+06 pystones/second
Adafruit CircuitPython 5.0.0-beta.4 on 2020-01-22; Adafruit Feather nRF52840 Express with nRF52840
Pystone(1.2) time for 500 passes = 0.683998ms
This machine benchmarks at 730996 pystones/second
short phoenix
Does anyone know when the Mu patch for Mac is going to come out that fixes the issue of text not being the right color, dark mode not working...?
slender iron
@short phoenix are you using the latest version? one was released over the weekend
solar whale
@short phoenix itβs out now.
short phoenix
Okay, thank you
tulip sleet
@slender iron uhhh..., in the nrf port, we have this:
void common_hal_mcu_disable_interrupts() {
}
void common_hal_mcu_enable_interrupts() {
}
and those routines are called various places. That might explain some issues ... oy
no reponse needed; just π€¦
slender iron
π
umbral ice
Hi all
Can I have some navigation help in the code base? I'm slowly working on rotaryio for the STM32 feather, but a total newbie to C and the circuitpython code structure. I see pertinent looking files in .../ports/stm32f4/common-hal/, is that the correct location?
tulip sleet
@umbral ice shared-bindings contains the Python interface to the native modules. shared-modules contains port-independent code that implements various modules. port/<someport>/common-hal contains port-specific implmentations of native modules. So you would add ports/stm32f4/common-hal/rotaryio
umbral ice
@tulip sleet Thank you. What is the equivalent workflow to python REPL?
tulip sleet
i'm not clear on what you're asking
umbral ice
Incremental guesses at seeing if I have the code right. I have the build environment set up. Do I just compile and see if it builds?
The iterative/testing process is what I'm unfamiliar with
tulip sleet
yes, you build it, get a .uf2 and load that onto the board and try it. But we use J-Link debuggers and gdb to debug issues all the time. You could use an STlink with the stm32
here is an old guide that talks about samd21, but the J-link part is relevant to stm32 as well: https://learn.adafruit.com/debugging-the-samd21-with-gdb
we don't use OpenOCD, in general
umbral ice
<reading... many thanks for the resources>
tulip sleet
what is your host computer?
https://learn.adafruit.com/circuitpython-samd-debugging-w-atmel-studio-7 talks about Windows, but I essentially never use Windows. I use Linux for development, Scott uses MacOS, but in a command-line way
umbral ice
Linux, it's been a while but nice to return
tulip sleet
do you have an STlink or J-Link?
or if you have one of the supported STM dev boards, they have integral STlinks, so that would work
umbral ice
I don't have ST/J-Link. It looks like the J-Link is cost prohibitive for me EDU mini is affordable.
One of the supported dev boards? I have one of the feathers, is that what you're asking?
tulip sleet
@umbral ice No, I thought perhaps you had one of the STM Discovery boards.
The EDU Mini is a good choice because it works on STM and SAMD and nRF and many others.
You could get an STlink cheaply as well but it's meant for STM only.
(sri my attn was diverted)
you will also need https://www.adafruit.com/product/2743
quartz raptorBOT
tulip sleet
because the Feather doesn't have an SWD connector on board. You'll need to solder some jumpers to pads on the bottom
umbral ice
quartz raptorBOT
tulip sleet
though i am not sure the 752 box header will fit. However, I have trimmed off the end of one on another board to make it fit against the pins.
umbral ice
A debugger is not the same thing as a logic analyzer, right?
tulip sleet
The box headers are a bit tricky to solder if you don't have SMD soldering experience. Use plenty of flux and feed new solder in. My success rate is not 100% on those. Soldering to the SWDIO and SWCLK pads near the SD socket would be easier.
no, not at all. a logic analyzer can be useful to check pin states on output. I use a Saleae sometimes, but I use gdb with the J-Link almost every day, and the Saleae every month or two
umbral ice
OK. I'm a chemist by training, some data science after college, now python and automation are part of my duties
tulip sleet
sounds like you are leaping into the frying pan a bit π
umbral ice
RE: SWDIO and SWCLK pads, would I only need those two?
tulip sleet
you need to also connect GND and the 3V pin to the 2743 SWD adapter above. Potentially RESET too, but that seems to be unnecessary nearly all the time. Do you have a breadboard and jumpers?
umbral ice
Yes. That is much easier than the SMD, and more forgiving for breadboard use!
tulip sleet
this shows an example of jumpers to the SWD adapter: it's for a different Feather board, but the idea is the same: https://learn.adafruit.com/how-to-program-samd-bootloaders/feather-m0-m4-wiring
Often I don't solder actual wires to the pads, but just a little piece of header, like this:
(looking for picture...)
then i use female jumpers on the header pins
but most people use wires
umbral ice
Either would work for me. I have a little breadboard dedicated to the project, I can add the debugger breakout
umbral ice
Thanks! This has been a good start
obsidian compass
This is surprising :
I have a few unpacked Particle Xeonon boards
Guess will now use em all with CircuitPython xD
solar whale
Yes -- same thought here. I was looking forward to learning more about mesh, though.... oh well - the Xenon works will with CP.
obsidian compass
@solar whale yeah I wanted to learn more about mesh too, especially the openthread implementations.
Is BLE supported under CP for xenon?
rotund basin
Is there a cat-m or lte solution with ethernet? I have some remote modbus devices I need to IoT π
idle owl
@gilded cradle Ping me if/when you're around please. Thanks!
solar whale
@obsidian compass yes, BLE is supported on the xenon. Same as the other nRF52840 boards
obsidian compass
That's amazing, now I just need to source a Jlink debugger
idle owl
@tidal kiln Maybe you know: how do I test something that is adding code to use bitbangio if busio SPI is not available? which is to say, what uses bitbangio?
And not busio.
tidal kiln
to test is busio.SPI is available - could try import and catch if it fails
idle owl
Right, the code exists, how do I test it?
Someone put in a PR to add this to DotStar
and I have no idea how to test it
Is there a particular microcontroller I need to use or..?
tidal kiln
let me look at PR...
idle owl
yeah......
so I'm wondering if this is even needed?
The original PR comments that it's way faster on every board, or some such.
Which I have no way to comment on
HW SPI is fast.
so?
tidal kiln
yah, typically it's the other way around. HW is faster than bitbang
idle owl
hmph.
tidal kiln
maybe something in blinka land
idle owl
Bleh.
tidal kiln
gemma m0?
idle owl
No HW SPI there? hmm
tidal kiln
chip has it, but not broken out, since there are so few pins
idle owl
hmm fair enough
https://learn.adafruit.com/circuitpython-essentials/circuitpython-dotstar doesn't mention that.
Did I miss something there then?
I probably didn't test the example on a Gemma...
Ah bugger, they're out of stock too.
I left a comment on the PR, no idea if the submitter is still active or not.
tidal kiln
yah, probably worth pinging submitter to try and get more context
idle owl
That's what I did.
tidal kiln
the current dotstar lib essentially falls back to a bingbang spi already:
https://github.com/adafruit/Adafruit_CircuitPython_DotStar/blob/master/adafruit_dotstar.py#L241
idle owl
The original PR said that bitbangio was faster than _ds_writebytes...
whatever that means.
So... we'll see.
tidal kiln
well, the above is bitbanging in python, so i can see how it would be faster to use bitbang from the CP firmware, i.e. bitbanging in C
ionic elk
@tulip sleet are you online?
tulip sleet
yes, just got on after doing some paperwork "errands"
ionic elk
got a second for a quick bootloader conversation?
tulip sleet
sure
ionic elk
amelia?
tulip sleet
@ionic elk ok back sorry
idle owl
@tulip sleet A question in #general-tech that you may be able to answer about BLE and CircuitPython, whether we've done a specific example.
idle owl
@tidal kiln Their response: @kattni i needed this on a trinket 5V with a WS2801 strip, where the Hardware SPI was already used by something else bitbangio was much faster than the _ds_writebytes() implementation ... how are they using CircuitPython then?
gilded cradle
Hi @idle owl, I'm around now
idle owl
@gilded cradle Excellent timing. What hardware is needed to test these two PRs? https://github.com/adafruit/Adafruit_CircuitPython_RGB_Display/pulls
I'm setting FoamyGuy up to test PRs, and he tried but gets an error with the hardware he has, and I'm not sure if the error is a problem with the PR or him having the wrong hardware.
gilded cradle
Ah, ideally it should be tested on each of the displays that RGB Display supports and a Raspberry Pi. It could be done with breakouts and/or PiTFTs. Let me get a list of supported displays
gilded cradle
Yeah, eventually. I keep having high priorities fall in my lap. The reason I haven't tested it yet is because I think it will take a while to properly test. I think those PRs were focused on one display and affect the other displays as well, so I will probably end up coming up with a bunch of changes.
idle owl
@gilded cradle Ok, I'll leave those to you then. Sounds like it's way more involved than it looked like at first glance.
gilded cradle
Yeah
idle owl
Right on. Thanks!
gilded cradle
yw
idle owl
@tidal kiln I'm not missing something with the Trinket 5V am I? There's no way that lib is being used on a 5V trinket?
tidal kiln
if trinket 5v = PID1501, then yah, no way it's running CP
same for PID2000, aka the "pro" trinket
tidal kiln
lib wise, yah, there is this:
https://github.com/adafruit/Adafruit_CircuitPython_WS2801
idle owl
@tulip sleet The question is from SilverAdept at 11:13amET. I don't know the answer, but you might.
@tidal kiln Yah I commented about that too.
ionic elk
oh @tulip sleet one other question - do you have a quick command to get the jlink to just flash a file, without opening the gdb server? All I see in the manual are UI tools
tulip sleet
@idle owl ni will look now
idle owl
Thanks!
obsidian compass
Can someone do a quick typecast an int to hex and check the type of hex variable for me?
>>> type(a)
<class 'int'>
>>> b = hex(a)
>>> type(b)
<class 'str'> ```hex() returns the hexadecimal string representation of an integer
ionic elk
oh, I assumed you were away from the computer
tulip sleet
there is no diffeerence between a decimal and hex integer. It's all binary under the covers
obsidian compass
Thank you @ionic elk , I wanted to double check π
tulip sleet
you can read and write an integer as a decimal number, a hex number, a binary number ,etc. but it's all the same number
they're all ints
obsidian compass
@tulip sleet makes sense , I need to tweak the code
ionic elk
I use https://www.rapidtables.com/convert/number/decimal-to-hex.html for long ones when I'm lazy
but obviously if you're already in python just use hex()
obsidian compass
@tulip sleet I have a hex string "4800ff"
I need to append 0x
should I use "0x" + string
?
or hex(string)
tulip sleet
do you want to print it or convert it to an int?
obsidian compass
I want it in a format which I can use it with "pixels.fill(val)"
which is int I guess
tulip sleet
you can give a base and use int:
>>> x = "4800ff"
>>> int(x, 16)
4718847
@tulip sleet thank you! trying this now
tidal kiln
or are you just wanting to do this?
val = 0x4800ff
pixels.fill(val)
is it a string you read in from elsewhere or did you write it yourself in your program?
tidal kiln
then what dan said, but curious why you are dealing with a string?
tidal kiln
ok, what dan said then
tulip sleet
it's confusing, and languages like javascript don't help. like, how horrible is this?
$ node
> "3" + 4
'34'
> "4" + 3
'43'
> 3 + "4"
'34'
> 3 - "4"
-1
>
^^ i think i saw a talk on that