review only. it's pypi setup only.
#circuitpython-dev
idle owl
@tulip sleet if you get a minute: https://github.com/adafruit/Adafruit_CircuitPython_TrellisM4/pull/7
pastel panther
@slender iron continuity seems fine (cut open a cable, measured to the pins on the ic). No shorts as far as I can tell. sense pin isn't connected to anything.
idle owl
Can this chunk of code be accurately described as "iterating over the NeoPIxel grid using the width and height values"? python for x in range(trellis.pixels.width): for y in range(trellis.pixels.height): pixel_index = (((y * 8) + x) * 256 // 32) trellis.pixels[x, y] = wheel(pixel_index & 255)
Context: https://github.com/adafruit/Adafruit_CircuitPython_TrellisM4/blob/master/examples/neopixel_trellism4.simpletest.py
pastel panther
how would the clocks be messed up?
tidal kiln
@idle owl who's the audience? might be ok to just say "looping"
idle owl
@tidal kiln People using the TrellisM4 library. It's going into the board guide. A page on using the lib.
tidal kiln
use either one
idle owl
ok. I wanted to make sure I was correct in my understanding of what iterating is. I looked it up, and it seemed like I was, but I wanted to verify with the example I would be using.
tidal kiln
for this, same thing, and i'm not even really sure when the difference in semantic minutiae matters
idle owl
Ok
exotic pumice
I think you're right, but iterating doesn't necessarily have to be in a loop
like a generator is kind of iterating too
I think
tidal kiln
@idle owl that PR looks boiler plate, did a quick copy-pasta scan, travis is happy, want me to merge?
idle owl
Yes please
pastel panther
@slender iron I have to admit defeat, have a little breakdown and then go to work π
thanks for your help. at least I know it's running code
slender iron
ya, I'm with you. I have a similarly frustrating issue
tidal kiln
@idle owl done
idle owl
Thank you!
pastel panther
good luck with yours, ttyl
slender iron
thanks! you'll figure your issue out with fresh eyes
idle owl
@tidal kiln @exotic pumice thanks for the input. I used iterate in a general explanation of what to use width and height for, and then simpler wording in the specific explanation of the example.
tough flax
@tidal kiln - yup, I saw that
uneven yarrow
I'm running into an issue with circuit python, the gemma v2, and capacitive touch. When I set up my gemma for cap touch it works fine when I touch the pad directly or when I alligator clip into something like fruit, but when I try to attach to some conductive fabric I can't get the touch to respond
tidal kiln
@tough flax have you experienced similar issues with i2c pull ups?
uneven yarrow
and I've seen some tutorials that say to use a resistor to control the sensitivity of the touch, but I thought the circuit python cap touch was self regulating
tough flax
Not with that sensor
uneven yarrow
(though now I'm wondering if that self regulation was on the playground express specifically ...)
tough flax
I've had other issues with them on the relay board
@tidal kiln did adding a 10K pullup solve it?
tulip sleet
@uneven yarrow When you first create a TouchIn object, the "calibration" is done by reading the current .raw_value, which assumes it's not being touched. The threshold for touch is then that value + 100
tidal kiln
@tough flax unknown. still working that specific forum issue.
tulip sleet
You could take some data on this by connecting the fabric, creating a TouchIn, and then looking at the .raw_value with and without being touched, and see what the difference is. You can set .threshold yourself if you want to tune it. But it might just not work very well with a piece of conductive fabric, or a long wire (like more than 2-3 feet, I've seen).
uneven yarrow
π fantastic! awesome yeah I will give that a try. Thanks @tulip sleet !
ah ok, this is what you're talking about: https://circuitpython.readthedocs.io/en/2.x/shared-bindings/touchio/TouchIn.html#touchio.TouchIn.threshold
ty. I remember reading it when I was building a cap touch jack-o-lantern last month, but I completely forgot about it since then π
idle owl
@uneven yarrow Make sure you're always viewing the latest documentation: https://circuitpython.readthedocs.io/en/latest/shared-bindings/touchio/TouchIn.html
APIs change π
uneven yarrow
(working on my second cap touch related project now)
oh! didn't realize I was looking at old docs, ty @idle owl
bookmarked!
idle owl
Always check to see that latest is in the URL, that's how you know. And it'll say "latest" down towards the bottom left in green.
I have a lot of 2.x in my history and sometimes forget and wonder why things are weird π
uneven yarrow
def. I'll make sure to keep an eye out for the version. thanks π
crimson ferry
Does anyone know why the M4 products (e.g., ItsyBitsy M4 Express, Feather M4 Express) don't have capacitive touch (or do they)? Is it a library issue or a hardware issue? Microchip site seems to indicate that the SAM D5X line has plenty of it, but I find no mention on Adafruit product or CircuitPython pages, and import touchio fails on M4 feather.
idle owl
@crimson ferry The way that it works was significantly different than on the M0 and so it didn't port at all. We haven't gotten to sorting it for the M4s as a result.
crimson ferry
Thank you. Sounds like it could be a possibility in the future for these boards.
gusty kiln
a quick one for review: https://github.com/adafruit/Adafruit_CircuitPython_HT16K33/pull/26
(re: that forum thread @slender iron posted earlier.)
idle owl
@gusty kiln Merged. Are you good to do the release?
gusty kiln
@idle owl procedure is just tag version with commit message as release notes at this point? haven't done one in a while.
idle owl
@gusty kiln Yeah, minor bump or whatever, and include the "to use this install the bundle, to install use pip, here are the docs" part too. Should be able to edit/copy/paste from a previous release for that if it was done before
gusty kiln
cool, on it.
solar whale
cool -- hooked up a Bluefruit SPI breakout to an ItsyBitsyM0 -- the "simpletest" works -- woohoo!
idle owl
Bundle add fail. bleh.
tidal kiln
ooo...i'm good at that...did i do it again?
idle owl
No, that's on me.
Tried to add to the bundle before releasing and then didn't bother to read what the error said thoroughly enough to realise that was the issue. Fixing it now.
prime flower
Just finished listening to the CPY Weekly, +1 on the live review session idea
idle owl
I guess I should have checked the fancy new Adabot output that would have told me that literally every lib I tried to add did not have a release instead of failing my way through Travis checks to figure out which needed it... π€¦
FINALLY
slender iron
@crimson ferry correct, its a software issue
timber mango
I was able to borrow from seesaw to get a single NeoPixel in the array on Trellis M4 operational .. in AtmelStart. ;)
https://github.com/wa1tnr/ainsu-CamelForth-d51-usb/tree/master/boards/trellis_m4_usb/src/cf/bsp
tidal kiln
@tulip sleet itsy eraser worked. thanks!
https://forums.adafruit.com/viewtopic.php?f=62&t=143286&p=707580#p707576
raven canopy
@idle owl was the need-release list accurate?
idle owl
@raven canopy yeah
idle owl
as far as I can tell anyway, heh. I didn't check all the empty repos. But now the non-empty ones all have initial releases.
tough flax
So... just thought I'd share. I've spent the week very frustrated. I put a reverse proxy ("edge server") in place in front of my day job's primary servers. I've done this many times to speed things up in my former clients, but this was on a service (and company) that I own, so it was "all me".
And it _just_didn't_work. I mean, everything got slower! It was insane! Server was sitting at 0.1 load, no IO, no network traffic, and yet when we went through the proxy, it crawled. Every 100th request or so would totally time out.
I finally pulled the plug and tried other approaches to fix our growing pains.
Today, I tried to reuse the edge server for a temporary project and downloaded a file from github... it ran at 120K/sec! After a call to the colo, it turns out that the port I had plugged the new server into had throttling set to 1MBit/sec. That was it!
We're back on track and I'm so happy I had to share π
Sorry to be off-topic, but I've really appreciated having someplace with competent people who helped me this week π
manic glacierBOT
cedar beacon
do any circuitpython boards have wifi?
solar whale
@cedar beacon The only wifi support in CircuitPython a t this time is for the ESP8266.
cedar beacon
π
marble hornet
hey, i'm working on a project and want to revert my aster to a previous commit, i tried googling but the git bash is giving me this error ```jonah@DESKTOP-39O84BI MINGW64 /d/system (master)
$ git push origin master
To https://github.com/TG-Techie/TG-US.git
! [rejected] master -> master (non-fast-forward)
error: failed to push some refs to 'https://github.com/TG-Techie/TG-US.git'
hint: Updates were rejected because the tip of your current branch is behind
hint: its remote counterpart. Integrate the remote changes (e.g.
hint: 'git pull ...') before pushing again.
hint: See the 'Note about fast-forwards' in 'git push --help' for details.
jonah@DESKTOP-39O84BI MINGW64 /d/system (master)``` after i ran:
git reset --hard
git pull
**i edited a file and added a comment**
git add --all
git commit -m 'who would cross the bridge of death etc'
git push origin master```You can push -f, or you can use git revert, adding an explicit rollback commit
Push -f probably makes sense in this car. Donβt do it if anyone else might have pulled from the github repo.
tulip sleet
βforceβ. Usually implies itβs dangerous
tulip sleet
π you can mess up the history so other people's repos are out of sync
marble hornet
this: schedule_secondary_terminal_read just popped us while using mu and a cp device. can i ask what is it?
meager fog
@umbral dagger hey dave. just checkin' my guidezchedule today - are you planning to submit one? totally no pressure!
umbral dagger
I have a few in the works. One on extending CP is nearing "ready to submit".
meager fog
@umbral dagger just so mike and i know our day - is it likely to be submitted in the next 3 hours? (again totally ok if not!)
ruby lake
ok, finally solved that mux address issue
uneven yarrow
@tulip sleet Thanks for your help yesterday. I got my MVP working this morning π https://vimeo.com/299450215
tulip sleet
@uneven yarrow Watched the video: very nice project! You might want to try the conductive fabric in some jacket or sweater of your own. The fabric will be close to body capacitance when it's worn, so you might need adjust the sensitivity in some way
@slender iron I think I am going to accept arturo's PR as is and then modify it from there ("we guarantee the API will change").
uneven yarrow
ah cool. Yeah I'll give it a try. Based on the testing I did this morning I figured I'd need to adjust the threshold as I add it to the hoodie, but good to know that it's likely going to change
slender iron
@tulip sleet sounds good to me
tulip sleet
kk, I'll get it up to date and submit a PR
coarse topaz
I'm using a DS3231 and compiling the latest version from git- I get RuntimeError: SDA or SCL needs a pull up when I try to initialise the I2C object with i2c = io.I2C(board.SCL, board.SDA)
tidal kiln
is the DS3231 actually attached ?
coarse topaz
π€¦
tidal kiln
@coarse topaz i can see how that message isn't really helpful if you don't understand the inner workings of I2C
@tulip sleet @slender iron should that message be appended with something like "check wiring" or "is device connected"?
slender iron
@tidal kiln sure! feel free to refine it
tidal kiln
d'oh. that's what i get for asking. π
@coarse topaz what do you think? what would be more helpful? keep in mind it needs to be as terse as possible.
pastel panther
@slender iron I don't suppose you have any thoughts on what hardware wise could mess up the usb clock?
tulip sleet
@pastel panther does your board have a crystal? this is m4, right? Which board of ours does it most closely resemble. I'd say this is more likely software setup
pastel panther
( on a samd51 board)
it does have a crystal, probably closest the the feather m4 express
tulip sleet
@tidal kiln would "no pullup found" be clearer than "needs a pullup"?
but we are running xtal-less because the DFLL is messed up, so any setup code should resemble the metro m4
or feather m4
pastel panther
I don't know if you caught the earlier discussion, but uf2-samd also can't get usb to work
pastel panther
51G
strange thing is it's the same basic uf2 and CP config that I've used successfully with other boards.
tidal kiln
@tulip sleet maybe. but also, some text that doesn't require knowing what "SDA", "SCL" or "pullups" are.
tulip sleet
maybe "no I2C devices detected (no pullups found)"?
@pastel panther so it's like an itsy (51g)
coarse topaz
That's much clearer
tulip sleet
but crystal stuff is broken due to DFLL errata; we never tried a 51g with a crystal, I think. is your uf2 board copied from itsy?'
slender iron
@pastel panther I usually check the clocks by outputting a PWM signal with a known frequency
pastel panther
@tulip sleet I don't recall at this point
@slender iron ok, I guess I can try that
tulip sleet
there is some commented-out code that outputs clocks on pins in supervisor/port.c
idle owl
@tidal kiln Note about initial PRs to empty repos: when you fork it, you need to delete the filler README.md otherwise it defaults to using that one and not the fancy one you added.
tidal kiln
@idle owl did i not linux correctly again? i thought i had done that π¦
idle owl
It doesn't get replaced when you copy because it's a different filename. So it's not a linuxing fail, it's an extra step I think. It was on BMP3XX. No worries. Simply wanted to make sure you knew.
tidal kiln
oh...hmmm. i thought there was a README.md in the cookiecutter stuff - so it would replace.
tidal kiln
sweet. spa treatment bonus!
idle owl
π yes
tidal kiln
"repos leave clean and refreshed!"
tulip sleet
@idle owl @tidal kiln would you like to switch to README.md's in the cookie-cutter in the long run?
tidal kiln
i'm fine with whatever. defer to @idle owl
idle owl
Really the only thing that would do is eliminate the step of deleting the readme.md in the initial PR.
Which I guess might be worth it.
tulip sleet
i just mean do you prefer RST or markdown when writing a README?
tulip sleet
me too
idle owl
I don't know what the badges would look like in markdown
(looking at a readme.rst right now)
timber mango
@pastel panther where is your code?
idle owl
Welp, figured out how to do the badges in Markdown.
I think it's worth looking into.
@raven canopy When you're around, I want to discuss changing the README in cookiecutter to markdown instead of RST.
pastel panther
@timber mango sitting right here
taps laptop
Haven't checked it in yet but the CP32-M4 board files are pretty much what I used for the new board, but with different pins
timber mango
$ ag PINMUX_P.*GCLK_IO0
samd51g19a.h
128:#define PINMUX_PA30M_GCLK_IO0 ((PIN_PA30M_GCLK_IO0 << 16) | MUX_PA30M_GCLK_IO0)
132:#define PINMUX_PA14M_GCLK_IO0 ((PIN_PA14M_GCLK_IO0 << 16) | MUX_PA14M_GCLK_IO0)
136:#define PINMUX_PB22M_GCLK_IO0 ((PIN_PB22M_GCLK_IO0 << 16) | MUX_PB22M_GCLK_IO0)
samd51j19a.h
160:#define PINMUX_PA30M_GCLK_IO0 ((PIN_PA30M_GCLK_IO0 << 16) | MUX_PA30M_GCLK_IO0)
164:#define PINMUX_PB14M_GCLK_IO0 ((PIN_PB14M_GCLK_IO0 << 16) | MUX_PB14M_GCLK_IO0)
168:#define PINMUX_PA14M_GCLK_IO0 ((PIN_PA14M_GCLK_IO0 << 16) | MUX_PA14M_GCLK_IO0)
172:#define PINMUX_PB22M_GCLK_IO0 ((PIN_PB22M_GCLK_IO0 << 16) | MUX_PB22M_GCLK_IO0)
Those are the pins that can output clocks.
G19A: PA30, PA14, PB22
J19A: PA30, PB14, PA14, PB22
This include file decides if the clock output is enabled on one of those pins:
// <q> Output Enable
// <i> Indicates whether Output Enable is enabled or not
// <id> gclk_arch_gen_0_oe
#ifndef CONF_GCLK_GEN_0_OE
#define CONF_GCLK_GEN_0_OE 0
#endif
```from `config/hpl_gclk_config.h`That's Atmel Start based. Dunno what other programming systems use.
raven canopy
@idle owl iirc, rST is used over Markdown so that Sphinx renders correctly.
idle owl
@raven canopy ah ok. nm on that one then.
@raven canopy not sure I'm prepared to fight Sphinx to get it play nice with markdown.
I guess on the plus side I now know two ways to use Markdown to make an image with alt text and a target link.
raven canopy
[!alt text] right?
raven canopy
idle owl
oh. hmm.
raven canopy
note: would require change to docs/index.rst and docs/conf.py as well for every repo. might be able to adabot patch it though
idle owl
Yeah I figured it would require a massive redo of Things β’
raven canopy
All the Things! π
idle owl
Though it wouldn't matter in existing repos if we didn't change the existing READMEs right?
Or is that something Adabot could do
raven canopy
No, wouldn't affect existing. Would create a rift in the standard. Haha
idle owl
Also, can I say, Adabot is pretty much amazing.
raven canopy
She is. And getting amazinger. I hope.
idle owl
Could Adabot update the readmes to .mds? Or is that way ridiculous.
timber mango
@pastel panther this should run on your board and demo USB CDC ACM adequately:
https://github.com/wa1tnr/CamelForth-SAMD51-SAMD21/tree/master/boards/trellis_m4_usb
pastel panther
@timber mango thanks for that; I'll give it a try when I get back to the bench
timber mango
that was my big achievement over the summer: USB CDC ACM in Atmel Start, for both M0 and M4 targets. ;) /a_whole_summer
clock foo is here:
obsidian compass
Hello everyone, my name is Ayan, I used to be active here few months back but was travelling for a while now not being updated with the progress on circuit python, but now I am back and ready to catch up and hopefully contribute here π
timber mango
Hi Ayan!
obsidian compass
Hi @timber mango
raven canopy
@idle owl we can test it. I imagine that a patch would work; just depends on how git would gen/handle the diff.
idle owl
Fair enough.
obsidian compass
The new trellis really caught my eyes in newsletter this week ..
Seems out of stock already
wanted to order one.
stuck elbow
those things often appear out of stock, as they are announced while they are being produced (I think)
obsidian compass
Ah! I see
stuck elbow
so you can subscribe and get a notification when they appear
solar whale
@stuck elbow received 2 Pew Pew 10.2 yesterday -- both up and running -- well done!
stuck elbow
@solar whale awesome!
solar whale
now I just need to find time to play with them!
stuck elbow
haha, I have the same exact problem
solar whale
with the ble WIP PR merged, I get ```jerryneedell@Ubuntu-Macmini:~/circuitpython_master$ git submodule update --init --recursive
error: Server does not allow request for unadvertised object b96950abf27229c2a3b1719d60691321246bfc94
Fetched in submodule path 'ports/nrf/nrfx', but it did not contain b96950abf27229c2a3b1719d60691321246bfc94. Direct fetching of that commit failed.
someone forgot to push?
slender iron
@solar whale I think you may need to sync your submodules
we changed to our own nrfx repo
travis passed so the checked in state is ok
solar whale
how do I sync submodules?
slender iron
(.env) /V/E/c/l/tinyusb (tinyusb_samd|β) $ git submodule sync
Synchronizing submodule url for 'hw/mcu/nordic/nrfx'
then git submodule update --init --recursive
solar whale
thanks -- happier now
slender iron
π
solar whale
so far so good ```Adafruit CircuitPython 4.0.0-alpha.2-135-gd08747d37-dirty on 2018-11-07; PCA10059 nRF52840 Dongle with nRF52840
help('modules')
main busio microcontroller supervisor
analogio collections micropython sys
array digitalio neopixel_write test
binascii framebuf os time
bitbangio gamepad pulseio uio
bleio gc random usb_hid
board hashlib storage
builtins math struct
Plus any modules on the filesystem
import bleio
dir(bleio)
['name', 'Address', 'AddressType', 'AdvertisementData', 'Characteristic', 'Descriptor', 'Device', 'ScanEntry', 'Scanner', 'Service', 'UUID', 'UUIDType', 'adapter']
solar whale
@tulip sleet still issues using USB to copt fiels to nrf52840 after ble active
tulip sleet
not surprised - I have not tested this, just reviewed the API. a lot of this code is from ubluepy. Did it have the same issue?
solar whale
but a simple scan works ```from bleio import Scanner
def display_scan_results(scan_entries):
for e in scan_entries:
print("ADDR: ", e.address)
print("NAME: ", e.name)
print("RSSI: ", e.rssi)
# Line break between record sets
print("")
Scan 1s for advertising devices in range
s = Scanner()
scan_res = s.scan(1000)
Display the scan results
display_scan_results(scan_res)
main things that will change coming up is that Device will probably be split back to Peripheral and Central again.
solar whale
ADDR: Address('b8:78:2e:19:70:42')
NAME: None
RSSI: -86
ADDR: Address('1c:e7:69:c0:22:d5')
NAME: None
RSSI: -86
ADDR: Address('45:4d:2c:4f:63:e2')
NAME: None
RSSI: -52
ADDR: Address('c8:69:cd:1c:5d:f8')
NAME: None
RSSI: -86
ADDR: Address('1c:e7:69:c0:22:d5')
NAME: None
RSSI: -82
what BLE devices are you scanning? I don't have many to test for.
solar whale
I have a bunch of Pcs and phones nearby
tulip sleet
so a lot of centrals, but not peripherals (or not peripherals only)
solar whale
@tulip sleet also trying on feather_nrf52832 -- so far - ok -- Control-C still does not work π¦
tulip sleet
same as before? not sure about that - I haven't tried it in a while
solar whale
yes -- has not been working for a long time -- works at REPL prompt, but wont interrupt a runnig script
ble_scan works same as on 52840 -- no hangup issues with USB -- yet
my other programs --not BLE related seem to be working as before -- SPI - SDCard stuff
@tulip sleet All of a sudden BLE is everywhere -- been having lots of fun with the new BluefrutiSPI library for the NRF51 devices. Works great!
tulip sleet
glad to hear! Lots of BLE is certainly a goal!
I have to test the SPIM3 bug again with the new code
solar whale
I'll be tied up tomorrow, but hope to have lots of time to poke/prod things this weekend. - My weekends now start on Friday π
idle owl
@solar whale congrats! π
solar whale
Thanks !!
umbral dagger
@solar whale Heh.. My weekend starts on days that end in Y.
solar whale
I'm getting there!
umbral dagger
The flip side is that everyday becoms a work day
solar whale
work is in the eye of the beholder
stuck elbow
and then where is the burnout?
umbral dagger
That's why I keep several projects on the go at all times.
Plus video games.
A small, fuzzy kitten also helps.
solar whale
It's a delicate balance -- kittens or puppies do help
stuck elbow
sorry, that was against the coc
solar whale
π Still trying to get my dog to understand the return to "Standard time"
stuck elbow
DST is undefensible
solar whale
agreed - even my dog knows that
stuck elbow
good thing we are getting rid of it in EU, finally
solar whale
nice!
stuck elbow
this will be the best thing right after the metric system :)
solar whale
the dog is demanding dinner -- back later
stuck elbow
β bed
marble hornet
is the pyboard supported? it looks like the stm branch is mostly left alone
raven canopy
@marble hornet not in CircuitPython. MicroPython yes.
marble hornet
thanks @raven canopy
to anyone:
could someone help me learn how to sceen share during show and tell?
marble hornet
pinged
raven canopy
question: anyone ever had file writes truncated inside
try:
gen_list()
finally:
f.write("\n".join(list))
i'm losing like the last 6 lines...
EDIT: nevermind. text editor on ubuntu just refused to show the last few lines with the window maximized. weird...
pastel panther
heya @slender iron I'm going to spend some time tonight trying to dig a bit deeper into this usb issue but honestly I'm struggling to motivate myself to debug the clocks because I don't understand what I could have done hardware wise to mess them up
pastel panther
because you don't know what I've done or because hardware making clocks broke doesn't make sense to you either?
I mean it doesn't not make sense, but I don't understand it (beyond an external crystal being messed up)
raven canopy
@pastel panther are your changes up on github? i am far-from-expertly, but have studied the clock code and may be able to see something...
unless, you haven't messed with clock code...
pastel panther
nope
I haven't, just added a set of board files
I have the same problem with the bootloader as well
so it's not CP thats the problem
raven canopy
ahhh...well, i am of even less help then. π
pastel panther
I appreciate the offer regardless
raven canopy
π
slender iron
debugging hardware is hard
pastel panther
that's very true
raven canopy
are you getting any USB messages? or just crickets?
pastel panther
nada, as far as I can tell
dmesg doesn't show anything, nothing in system info, no mass storage
I might poke the data lines with my scope later just for kicks, but I don't think I'll be able to make any sense of it
tough flax
Wow. Just realizing how tight memory is going to be on this Trinket M0 KeySwitch... probably should have spent the extra $5 each on an Itsy M4
I might have to take Dan's fall-back approach & free adafruit_hid modules into a custom build & get rid of Math and stuff like the PIRkey
freeze (not free)
So I just realized something about Python. I imported KeyCodes in main.py... and then I imported it in a module file... it actually used twice the space!
Is there a way to share that import between the files memory-wise?
velvet badger
Anyone know if there is a way to solder an off switch onto a crickit?
upbeat plover
import gc # needed to keep memory nice
def write(f, what_to_write):
with open(f, "a") as fp: # open to add line to file
fp.write(what_to_write) # write
fp.flush() # what does this do, hope it helps????
def read_in_chunks(f, size=64):
while True:
chunk = f.read(size)
if not chunk:
break
print(chunk)
yield chunk
def copy(file_from, file_to, size=64):
with open(file_from) as f:
for chunk in read_in_chunks(f, size):
write(file_to, chunk)
gc.collect() # this is very importaint without it the memory fills
# larger size will use more memory per chunk but increases speed greatly
read_in_chunks("test.txt", size=64)
copy("test.txt", "test2.txt", size=64)
gc.collect()
print(gc.mem_free())
print(gc.mem_alloc())
read and copy large files, with a set amout of RAM size
pastel panther
@velvet badger Doesn't it already have a switch?
velvet badger
A very tiny switch. I need one that's external.
tough flax
@velvet badger If you're using the Crikit Feather, you can wire between GND and EN - that will disable the Feather's 3.3v regulator which will stop your code. I will not stop powering motors, etc.
tough flax
Has anyone written a guide on how to freeze in a module in a custom build? I basically can make my code work if I don't import adafruit_dotstar, but I'd really like to be able to give visual feedback
One option would be to freeze dotstar & hid into the build
The alternative I've come up with is to export the configuration in a JSON-ish (is there a python serialize in CP?) and only parse the file when it changes
slender iron
@tough flax it shouldn't use twice the memory. How are you importing it?
tough flax
Perhaps not twice, an extra 1K each time
?
from adafruit_hid.keycode import Keycode
HEY! π mpy-cross got me just under the limit!
@slender iron - ignore me
I was importing the KeyboardLayoutUS as well when I ran that test
slender iron
hrm
tough flax
The only difference in size is the code for the import line (50bytes)
My bad
I also wasn't running gc.collect() before mem_free()
I'm just finding it very tight in the M0 Trinket
But mpy-cross will probably get me just where I need to be
with about 500bytes left π
You can't mpy compile main.py can you?
Just to catch you up on what I'm trying to do (@tulip sleet is doing something similar & we're collaborating)
I have a config file like this:
debounce=yes
bouncetime=.1
repeat=yes
repeattime=1.0
color=#555555
[common]
color=#FF0000
1 SPACE
2 ENTER
3 LEFT_CLICK
4 RIGHT_CLICK
5 LEFT_CLICK LEFT_CLICK
[mouse]
color=#00FF00
1 MOUSE_LEFT
2 MOUSE_UP
3 MOUSE_RIGHT
4 MOUSE_DOWN
5 LEFT_CLICK
[browser]
color=#0000FF
1 SPACE
2 TAB
3 SHIFT+TAB
4 PAGE_DOWN
5 PAGE_UP
I want to have "modes" for what the KeySwitches do (those three modes are pretty common) & some settings global and at each mode
The lines can have 'chords' too (like
1,5 CHANGE_MODE
The file would be on the Trinket and be editable by end users
Code to parse it now works (yay) and I am activating when the switches are pressed
when I tried to light the dotstar it barfed on memory
Not sending HID yet (but that should work)
After mpy-cross I have 1552 bytes left
So, my only remaining question before I crash is ...
Is there anything like pickle or some other serialize in CP? @slender iron
marble hornet
can i override the permission denied in the write function without unmounting the cp drive ?
tulip sleet
@tough flax I made you a trinket build with hid frozen in. See your dmβs. Didnβt read here first.
@marble hornet only if you use remount in boot.py. See a data logging learn guide for details.
@marble hornet e.g. https://learn.adafruit.com/cpu-temperature-logging-with-circuit-python
meager fog
@tidal kiln hey0
tidal kiln
@meager fog hey
meager fog
@tidal kiln basecamp's down - if ya have a todo list of sorts i can paste it here
tidal kiln
only todo currently on there was for the few wrap up things - cp libs, feedback servo guide update
are you looking for new todo?
tidal kiln
@idle owl ran into a blinka issue - do you take those also?
@slender iron trying to figure out how to stop using a servo after it's been created. looks like continuous has a deinit:
https://circuitpython.readthedocs.io/projects/motor/en/latest/api.html#adafruit_motor.servo.ContinuousServo.deinit
but the regular servo doesn't?
tidal kiln
ok. so i was looking for the right thing?
just to make sure i was reading how it's suppose to be done correctly
tulip sleet
@tidal kiln the deinit() doesn't really deinit, it just sets the .fraction to 0. The object to deinit would be the PWMOut that you passed in when you created the Servo or ContinuousServoc
I'm not sure why there's a deinit() in Continusous Servo
tidal kiln
thanks. was wondering about that approach. thought maybe the servo one did that plus other stuff.
tulip sleet
maybe you or I should file an issue
there's with support in ContinuousServo.
DCMotor also has a deinit() that just stops the motor without actually deinit-ing the PWMOut.
tidal kiln
here's more context...
trying to make a CP version of this:
https://learn.adafruit.com/analog-feedback-servos/servos-as-input-devices
so was looking for CP equivalents for attach and detach
those are needed for the record part, so user can move servo without it fighting back
tulip sleet
I think if you just set .throttle to zero, that would do it. The API is reallly pretty different.
tidal kiln
that's only on continuous
tulip sleet
hmm, yes, it looks like there shold be a turn_off() or release() or something that sets duty_cycle to 0
right now there is no straightforward way to do that
tidal kiln
ok, so i wasn't overlooking something.
like you said - for now, could just cheat and directly use the PWM object
tulip sleet
yes, though that's kinda messy ("leaky abstraction")
idle owl
@tidal kiln Depends on the issue, but either way you should file it on the Blinka repo probably.
tidal kiln
@idle owl it was for this:
https://learn.adafruit.com/neopixels-on-raspberry-pi/python-usage
i was messing with that this morning. i had to pip install rpi_ws281x also to get neopixel to work
idle owl
hmm. Ok, file an issue on the NeoPixel repo for that one. I'll talk to Brennen about it.
tidal kiln
is it an issue for that repo because one of the requirements files should've taken care of that?
vs. a blinka repo issue?
gusty kiln
@tidal kiln if you install it in a fresh venv, does it still not pull in the rpi_ws281x dependency?
tidal kiln
unfortunately, haven't done much with the venv yet. but one thing i did do different from guide page - i didn't run with sudo
for the initial neopixel lib install
gusty kiln
i'll poke at it.
tidal kiln
@tulip sleet can you please file the issue for the servo stuff? i think you're more dialed in on it than i am.
tulip sleet
sure
tidal kiln
thanks!
@gusty kiln it's quite possibly me. i wasn't following the guide directly. let me know if you need any more deets on what i did. thanks for looking into it.
idle owl
@tidal kiln Something seems odd on @gusty kiln's end as well.
tidal kiln
@idle owl @gusty kiln FWIW, it was from working this forum post, which may be Yet-Another-Issueβ’
https://forums.adafruit.com/viewtopic.php?f=60&t=143322#p707991
gusty kiln
@tidal kiln what pi do you have?
tidal kiln
zero w
gusty kiln
i bet this is the issue: rpi_ws281x>=4.0.0; platform_machine=='armv7l'
idle owl
ahhhhh
tidal kiln
and now that you ask, rings a bell, isn't blinka advertised for pi 3 (and maybe 2)? goes to look.....
oh. no. it's ok with zero. it's just the much older pi 1's.
tulip sleet
@tidal kiln filed two issues on Motor
tidal kiln
@tulip sleet awesome. thanks!
tidal kiln
@meager fog yep. appears so. over to basecamp then?
meager fog
ill add it, no rush
gusty kiln
idle owl
@tidal kiln Can I get your opinion on something? Does any part of this seem like something that should get added to the CircuitPython Troubleshooting page? https://forums.adafruit.com/viewtopic.php?f=48&t=143343 I feel like not really except maybe something about "Errors in the Serial Console: Make sure the serial console window is tall enough to show the errors or it may appear that no error has been shown" or something worded better than that. And even that I'm not sure is worth adding. But you see support stuff way more than I do so I'm asking for your opinion.
tidal kiln
@idle owl i'd say add it. with a screen cap and some arrows pointing at the scroll bar in the serial window. and big text that says "SCROLL BACK AND LOOK UP THERE ^^"
its not the first time people have been tripped up by that UI quirk. esp. since the default height seems to be pretty slim.
idle owl
hah. Ok fair enough.
gusty kiln
@tidal kiln so i think the zero is a bcm2835, right? i.e., same or close to same ARM chip as a pi 1. testing on a pi 1, at any rate:
Python 3.5.3 (default, Sep 27 2018, 17:25:39)
[GCC 6.3.0 20170516] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import platform
>>> platform.machine()
'armv6l'```so i guess if the neopixel code will work on those older systems, we probably want to change that restriction.
tidal kiln
pi@raspberrypi:~ $ cat /proc/cpuinfo
processor : 0
model name : ARMv6-compatible processor rev 7 (v6l)
BogoMIPS : 697.95
Features : half thumb fastmult vfp edsp java tls
CPU implementer : 0x41
CPU architecture: 7
CPU variant : 0x0
CPU part : 0xb76
CPU revision : 7
Hardware : BCM2835
Revision : 9000c1
Serial : 000000009ead533c
@tidal kiln And it did work once you installed the dependency?
as in there's not some hardware restriction that we were working around
tidal kiln
no. ran into permission issue trying to access mem.
idle owl
that's the sudo thing I think. It's required to access the peripherals needed for NeoPixel.
Did you try it with sudo later?
tidal kiln
yep. but since i hadn't installed with sudo (my bad, guide has you do that), the module wasn't available then
gusty kiln
looks like we don't support neopixels on other than 2 & 3, so i guess this is a non-issue but we should make the docs clearer about that.
meager fog
2, 3 and Zero
maybe Pi 1 isnt' supported
but...i think if Zero is, Pi 1 is - same chip
gusty kiln
if board_id == "raspi_2" or board_id == "raspi_3":
from adafruit_blinka.microcontroller.raspi_23 import neopixel as _neopixel
elif "sphinx" in sys.modules:
pass
else:
raise NotImplementedError("Board not supported")```i dunno if that means it wouldn't work, but it's written to blow up if you try anyhow...
idle owl
@tidal kiln Eh?
meager fog
oh we just haven't written in blinka support yeah
@gusty kiln we should do that
do you have a pi zero to check the board_id
tidal kiln
@idle owl yep looks good
gusty kiln
yeah, can do
tidal kiln
@idle owl in text that goes along with this, could also mention re-sizing window as an option
idle owl
good call, thank you
@tidal kiln https://learn.adafruit.com/welcome-to-circuitpython/troubleshooting#serial-console-in-mu-not-displaying-anything-18-13 Look good?
tidal kiln
@idle owl yah. even more info than i was expecting. awesome.
idle owl
Excellent thank you
river quest
Python on hardware! 03 https://youtu.be/h4QbMHmnsDc
prime flower
@idle owl Did you write a SPI scanner at some point?
idle owl
For which pins are HW SPI? It's on the SPI Essentials page.
manic glacierBOT
idle owl
@slender iron Can you take a look at this to make sure there's not some more in depth stuff that should be changed? If you think it's good to go, I'll merge it. https://github.com/adafruit/adabot/pull/27
prime flower
Thanks Kattni, found it
dapper falcon
@idle owl Thanks again for the debugging help. The web page looks great.
idle owl
@dapper falcon You're entirely welcome!
@dapper falcon We have an educator's role on our Discord server, if you'd like to be added. The purpose is to make you more identifiable to other educators in the event that you'd like to discuss methods or curricula or that sort of thing. There's no obligation associated with it, nor is there any obligation to be added. I wanted to let you know so you are aware and if you'd like to be added, we can absolutely do that!
dapper falcon
@idle owl Sure. Sounds good.
idle owl
Thank you! You'll show up to the right under the Educator list when you're online. π
slender iron
@tulip sleet I'm about to land from this giant leap and I'm hoping I won't squish you. (Getting close to a PR)
anyone know where ble.h comes from in the new nrf build?
raven canopy
i can't even find it... π€·
slender iron
got it, had to run bluetooth/download_ble_stack.sh
raven canopy
ahh...dependency.
slender iron
time to do one more cleanup pass
tulip sleet
@slender iron I think we don't overlap very much if at all
slender iron
its just a huge PR
tulip sleet
@slender iron bleio allows keyword args for non-optional args, e.g. UUID(uuid=777) is OK. The arg must be present, the keyword is optional. In a lot of our existing builtin modules, we don't allow that (e.g. TouchIn(pin=board.A0) is NOT allowed), even though the arg is named pin in the documentation. Do you have an opinion? This is kinda typical for MicroPython: it saves code and flash space not to match regular args against keyword names. In CPython, it always allows keyword names even for non-optional args.
i started taking out the keyword checking, but then thought I would ask you
slender iron
I like allowing kwargs because it makes calls more readable
I bias towards always using them
tulip sleet
so in the long run, we might add them to all the places where we don't allow them right now
tulip sleet
ok, I'll just put the code back that came from arturo and glennrub, not a big deal
tnx
slender iron
np, thanks!
tulip sleet
@slender iron approved descriptor PR
slender iron
yay!
idle owl
π
slender iron
thanks for the quick turn @tulip sleet. now I may be able to get the other PR passing on Travis
slender iron
At least its net negative: +2,223 β3,487
idle owl
yay!
raven canopy
negative is good! well done, @slender iron.
manic glacierBOT
lime trellis
quick question π i'm able to compile my own CP port based off the feather M0.
where do I set pin states at startup?
so chip selects and such aren't floating
raven canopy
@lime trellis i think you could do it on the CircuitPython side by using a boot.py, which would get run before code.py/main.py.
exotic pumice
anyone know the difference between interrupts SERCOM0_0 SERCOM0_1, etc.
samd21 just has SERCOM0, but samd51 has more options
the docstrings are no help
raven canopy
@lime trellis hmm. it may be due to the calls to reset_port() and/or reset_board() after boot.py is run. so...maybe put the initial pin settings in reset_board(). Feather M0 express example here: https://github.com/adafruit/circuitpython/blob/master/ports/atmel-samd/boards/feather_m0_express/board.c#L37
exotic pumice
also, what's the equivalent of SERCOM0_CORE in samd51?
raven canopy
@exotic pumice the 0_0, 0_1, etc are "pads" for the I/O routing. from the datasheet:
33.5.1
I/O Lines
Using the SERCOM I/O lines requires the I/O pins to be configured using port configuration (PORT).
The SERCOM has four internal pads, PAD[3:0], and the signals from I 2 C, SPI and USART are routed
through these SERCOM pads through a multiplexer. The configuration of the multiplexer is available from
the different SERCOM modes. Refer to the mode specific chapters for additional information.
note: i'm pretty sure there are typos in that text. there are only two pads per sercom
exotic pumice
so what if I just want an interrupt handler for the whole sercom?
like in samd21, it's just this https://github.com/adafruit/asf4/blob/d270f79aa16dd8fd4ae3b6c14544283dcb992e9c/samd21/include/samd21g18a.h#L109
do I do nvic enable on each pad?
raven canopy
yeah, for each pad. i don't think 21 had multiple pads per SERCOM (don't have the sheet on this PC)
exotic pumice
ok thanks
do you know what I mean by SERCOM0_CORE?
it's the IDR value of SERCOM0, but samd51 has no IDR
so instead of gclk.genctrl.write(id, blah, blah, blah) it's something like gclk.genctrl[id].write(blah, blah blah) I think, but I'm not sure where to get the id from
I guess I just have to figure out the array positions and define it myself
great, they're just called 0-11
π¦
exotic pumice
I think table 14-9 might be what I'm looking for
timber mango
@raven canopy definitely four pads per SERCOM
https://github.com/wa1tnr/ainsu-CamelForth-d51-usb/blob/master/boards/metro_m4_usb/include/pio/samd51j19a.h#L486
<< unmodified upstream file from Atmel Start
I think for example you can have DTR or DSR pins for the UART with the two 'extra' SERCOM pads.
(all four SERCOM pads could be routed to physical, external pins)
exotic pumice
it's actually 3
at least according to my svd-generated crate
it can't find SERCOM0_3
timber mango
$ pwd | cut -b26-99
circuitpython/ports/atmel-samd/asf4/samd51/include/pio
$ ag SERCOM0_PAD? samd51j19a.h | egrep PORT
489:#define PORT_PA04D_SERCOM0_PAD0 (_UL_(1) << 4)
493:#define PORT_PA08C_SERCOM0_PAD0 (_UL_(1) << 8)
497:#define PORT_PA05D_SERCOM0_PAD1 (_UL_(1) << 5)
501:#define PORT_PA09C_SERCOM0_PAD1 (_UL_(1) << 9)
505:#define PORT_PA06D_SERCOM0_PAD2 (_UL_(1) << 6)
509:#define PORT_PA10C_SERCOM0_PAD2 (_UL_(1) << 10)
513:#define PORT_PA07D_SERCOM0_PAD3 (_UL_(1) << 7)
517:#define PORT_PA11C_SERCOM0_PAD3 (_UL_(1) << 11)
Looks like (in general) that one of two pins (PA04, PA08 for example) can be routed to each SERCOM pad.
I could be way off base, here. ;)
Not just two pins; for example, SERCOM5:
$ ag SERCOM5_PAD? samd51j19a.h | egrep PORT
1266:#define PORT_PA23D_SERCOM5_PAD0 (_UL_(1) << 23)
1270:#define PORT_PB02D_SERCOM5_PAD0 (_UL_(1) << 2)
1274:#define PORT_PB31D_SERCOM5_PAD0 (_UL_(1) << 31)
1278:#define PORT_PB16C_SERCOM5_PAD0 (_UL_(1) << 16)
shruggo
raven canopy
yep...correction accepted. IO table (6.1) does have 0->3 references for each SERCOM. was thinking i only ever saw 0 & 1.
raven canopy
same here. laptop thinks so too... π
timber mango
I can't think of an application besides DSR and DTR for the USART - but somewhere I'd already encountered this. Very confusing to me. ;)
(BTW only SERCOM5 seems to support all four pads)
exotic pumice
oh, that could be why they weren't showing up in my pac
timber mango
(skuze me -- this is hard -- only SERCOM5 supports four different pins per PAD). haha
$ ag SERCOM5_PAD0 samd51j19a.h | egrep PINMUX
that'll return four lines. Anything but SERCOM5 there will return fewer than four lines.
fades
tough flax
Hi folks, just pushed a new version of the KeySwitch code: https://github.com/ATMakersOrg/KeySwitch Many thanks to @tulip sleet and others for their help in getting this to fit on the Trinket M0!
I'd love some feedback on this - the docs are not done and it's not structured right (should be in src, etc.) but if you look at settings.ini I think you'll get the way it's supposed to work. Right now, individual switches as well as multi-switch "chords" work sending either keystrokes or mouse movements (no mixing right now ala "shift-click")
The code should run on M4s with plenty of room, etc. but with the parsing code, getting it on a M0 is tight. The FrozenHID.UF2 is a custom build freezing HID and removing Math and other stuff like the PIRKey does to make everything fit. KEYSWITCH.UF2 has that build Plus the actual CIRCYUITPY filesystem with the config files etc.
Undocumented, you switch modes by holding switches 1+2 for > 1sec (configurable later)
slender iron
@lime trellis the board.c file can do any init you want and then in the mpconfigboard you'll want to add the pins to the mask so they aren't reset
@exotic pumice I'm not sure what you mean by IDR. where is it mentioned in the datasheet?
cedar beacon
What simple, low energy display would people recommend for the metro m0? Got 3 sensor readings to display
stuck elbow
one of the oled displays, probably
or you could use a 7-seg display and display the readings in order
hoary hill
Hi, how do I request a new Circuitpython library? There is library for MCP23008/17, the I2C GPIO expander. But I need a library for MCP23S08, the SPI expander. Is it possible?
marble hornet
@cedar beacon on the product pages for the .96 inch oled and the super tiny st7735r display they both have around a 20mA current draw (5mA diff) https://www.adafruit.com/product/931
https://www.adafruit.com/product/3533
cedar beacon
Thanks @marble hornet
marble hornet
are you using python ?
raven canopy
@hoary hill you can file an issue on the CircuitPython GitHub repo requesting the library. https://github.com/adafruit/circuitpython/issues
hoary hill
Thanks @raven canopy , I just filed the issue #110.
raven canopy
Ahh. You put it on the bundle. The "#110" threw me off; core repo is in the 1300s. π That should be fine.
marble hornet
i'm having an issue with exec
im running:
print(fmoop)```and getting
and eval is saying i have a syntax error but the exec('from '+path+' import ' + name + ' as fmoop') is just 'from system.programs import sys_bar as fmoop'
tulip sleet
@marble hornet maybe there's an error in the code you're importing? (as opposed to in those two lines)
marble hornet
i've ran the string as regular code and it works
solar whale
@marble hornet you need double ' for the inner quotes xec('from ''+path+'' import '' + name + '' as fmoop') not a double quote " -- two single quotes '
marble hornet
@solar whale what does two quotes do vs "
solar whale
not sure but - you need two f the same to tell it that it is not the terminator of the first '
it thinks 'from ' isa string ' from ' ' then it the second single quote is an inner string
solar whale
exec('from ''adafruit_rgb_display'' import ''rgb'' as ''bgr'' ') works for me
those are all single quotes
exec("from ""adafruit_rgb_display"" import ""rgb"" as ""bgr2"" ") alos works -- all double quotes
stuck elbow
ugh, use %
solar whale
what is the syntax for %
solar whale
Thanks!
stuck elbow
solar whale
BTW -- I was not advocating the method or style -- just trying to explain the "syntax error"
stuck elbow
the explanation is simple: eval only can run expressions, and import is a statement
the quotes have nothing to do with it
copper shuttle
hello everyone
I just installed Mu to work with my Gemma M0 but I can'T launch it?
I click on it and simply nothing happens
No loading, no nothing
Already restarted my PC twice and still nothing
Anyone got an idea?
slender iron
@umbral dagger I just read over your extending guide. I think you'll want to use shared-modules instead of common-hal because your libraries are not mcu specific
the structure is described here: https://circuitpython.readthedocs.io/en/latest/docs/common_hal.html
umbral dagger
Cool. I'll pull it back and make that change.
slender iron
thanks for writing it up!
idle owl
@tidal kiln You around? I'm setting up the MPL3115A2 to test it. Do SDWN or RST need to be wired to anything? Also do you happen to have a .mpy I could use?
tidal kiln
let me remember/check....
idle owl
There's no assignment for either pin in the simpletest.py so I'm guessing no. But there's no guide for it so I'm assuming.
tidal kiln
yah, they don't need to be hooked up.
idle owl
ok right on. Do you have an .mpy?
tidal kiln
looks like i was just testing with the .py
idle owl
ok.
ruby lake
hm, the Metro M0 Express pages need to explain that IORef is tied to 3.3v
tidal kiln
pretty easy to mpy-cross it if you want though
idle owl
I can do it, I figured if you had it, I would use yours
tidal kiln
i typically just try the .py to see if i can get away with it, then mpy-cross if not
tidal kiln
cause my dev setup isn't much more than cp and sync
slender iron
@ruby lake m0 or m4?
ruby lake
@slender iron m0, https://learn.adafruit.com/adafruit-metro-m0-express-designed-for-circuitpython/pinouts
I would add a line in the power connections section saying IORef is tied to 3.3v on the board, or something like that.
slender iron
Why would it be something else?
umbral dagger
@slender iron What about where I use atmel_start_pins.h ?
slender iron
why do you need it @umbral dagger ?
@ruby lake added it under other
ah, @umbral dagger use digitalinout internally like you do in the python
idle owl
@slender iron Are you ok with us setting up Sphinx to render Markdown and moving README.rst to README.md if we can sort doing it reasonably? Getting Sphinx to render .md is easy. It's the rest of it that involves "figuring out."
slender iron
I'd rather not because we can't use markdown inline last time I checked
idle owl
I don't know what you mean by "use markdown inline"
gusty kiln
i'll note that pypi markdown support also seems... kind of weak.
slender iron
in function docs and in the rst in the shared-bindings c
idle owl
ok.
I initially said nevermind on it but the Sphinx part was easy, so I reconsidered. But if it's going to complicate everything, no.
slender iron
I'd love to use markdown everywhere but we can't. so I'd rather use rst everywhere even though its not as friendly
idle owl
@tidal kiln I know you did everything with MPL before we talked about BMP, but one more reminder that when you fork the repo, delete README.md or it supersedes README.rst.
@slender iron Fair enough.
I have a repo where the docs are passing according to RTD (as in the Builds tab on the dashboard), but if you click "view docs" you get the not found page. And it says "docs failing" according to the badge on the GitHub repo.
And the docs badge also goes to a not found page.
I doublechecked that I added it with the right alias, and I did.
tidal kiln
@idle owl sry. yah. good chance i'll remember for like the next few then forget again. so many details. very glad you're on the job to catch these things.
timber mango
In July 2016 the Linux kernel project decided to transition from DocBook based documentation to reStructuredText and the Sphinx toolchain.[8]
cite
https://en.wikipedia.org/wiki/ReStructuredText
idle owl
@tidal kiln No worries. I forgot about it until I'd already merged and released, so I deleted it separately. I missed it too initially.
Also, nm, stable passed, latest did not, that's why it's showing failed because all the links are to latest. Oi.
winged grail
Does anyone have any good examples of synth programs for the neotrellis m5? I am just getting started in circuitpython, and so I would like to study the code as well as learn how to upload it to my board. Thanks sorry for a stupid question
sorry I meant a MIDI synth
idle owl
@winged grail That's not a stupid question! We're working on MIDI for CircuitPython right now, but it's not ready yet. As for getting started with CircuitPython, you'll want to start here: https://learn.adafruit.com/welcome-to-circuitpython
winged grail
awesome thank you so much!
slender iron
@winged grail do you want to generate midi or read midi?
winged grail
Generate, but I will take whatever anyone has, for learning purposes
slender iron
I'm hoping to add midi over usb soon. I think it works over uart now
tulip sleet
@slender iron You might or might not want to comment further here: https://forums.adafruit.com/viewtopic.php?f=60&t=143455
slender iron
will keep an eye on it and see what they want to do
exotic pumice
@slender iron, on samd21, the genctrl of each gclk has an ID register, but samd51 doesn't
exotic pumice
my samd21 code is
fn set_gclk_divider_and_source(
&mut self,
gclk: ClockGenId,
divider: u16,
src: ClockSource,
improve_duty_cycle: bool,
) {
self.gclk.gendiv.write(|w| unsafe {
w.id().bits(gclk.bits());
w.div().bits(divider)
});
I think what it's doing is "write divider to the gclk which was passed in"
the equivalent samd51
self.gclk.genctrl.write(|w| unsafe {
w.div().bits(divider)
});
self.gclk.genctrl is an array of 12 and I don't know which one to write to
because there's no id on the gclk passed in
idle owl
@solar whale thanks for testing my code!
solar whale
@idle owl no problem - thanks for providing it! 
tidal kiln
@idle owl i'm doing a major overhaul on a lib, it doesn't look like it has an example that complies with the naming boiler plate - is that critical to anything? worth making one?
idle owl
@tidal kiln there should be a sensor_simpletest.py example, so if there isn't one, you should create one.
tidal kiln
went with that. there's nothing in the automated stuff that expects sensor_simpletest.py?
idle owl
cookiecutter creates it by default and expects you to populate it, I thought. But maybe that was before we added that.
Not sure if Adabot checks.
Maybe Adabot checks to see if /examples is empty, but that's it.
slender iron
@exotic pumice thats a common difference between the 21 and 51. the 21 will make you write the id of the instance you want to change while the 51 just has you offset to the right one
exotic pumice
so is the offset still gonna be gclk.bits() ?
lime trellis
I know i've seen a write-up for building custom c-modules for CP. But for the life of me I can't find it.
Anyone know what I'm talking about?
slender iron
@exotic pumice I'm not sure what bits is. I don't know rust
exotic pumice
let me look it up
slender iron
@lime trellis there is this: https://circuitpython.readthedocs.io/en/latest/docs/common_hal.html
solar whale
hmm -- just pulled master and built for metro M4 -- boots ok -- installed modulel list seems a bit short ```
Press any key to enter the REPL. Use CTRL-D to reload.
Adafruit CircuitPython 4.0.0-alpha.2-141-g97bc95183 on 2018-11-09; Adafruit Metro M4 Express with samd51j19
help('modules')
main busio micropython storage
_os collections neopixel_write struct
_time digitalio network supervisor
analogio errno os sys
array >>>
and @umbral dagger is working on a guide
@solar whale I'll try it here. totally could be a bug
solar whale
@slender iron no way π
exotic pumice
yeah it looks like this is the equivalent https://atsamd-rs.github.io/atsamd/atsamd51j19a/atsamd51j19a/gclk/pchctrl/enum.GENR.html
slender iron
haha, there are always bugs
slender iron
pchctrl connects the clock to things
solar whale
I just thought I'd run the new USB stuff on as many boards as I could.
slender iron
I figured you would π
ideally they'll all have the same bugs π
@exotic pumice I'm having trouble helping you because I don't know the rust side. I could help better if you reference the datasheet instead
solar whale
ran some neopixels and I2C sensors on the M4 -- no issues.
exotic pumice
I think I got it figured out
lime trellis
thanks @slender iron ! I found what I was searching for: https://github.com/BitKnitting/wakey_circuitpython/wiki/Adding-Waking-from-Deep-Sleep-to-Circuit-Python and https://github.com/BitKnitting/wakey_circuitpython/wiki/Wake-Up-SamD21-Through-an-External-Interrupt
exotic pumice
but if you're curious look at gen in 14.7, under pchctrl
slender iron
glad you did @lime trellis
exotic pumice
bits is just the raw value of that field
slender iron
@solar whale I was considering redoing the output buffer for cdc already π
solar whale
did the new code go into esp8266 as well or just atmel and nrf?
slender iron
no esp because it doesn't have usb
solar whale
oh yeah -- small detail...
raven canopy
@idle owl @tidal kiln adabot doesn't currently check for simpletest. only that there is at least one file in the examples/ folder.
and cookiecutter does have examples/name_simpletest.py.
slender iron
@solar whale did you ever try help("modules") on nrf before? I bet it has the same problem
solar whale
@slender iron well -- pca10059 won't boot after loading new image π¦
slender iron
thats the dongle?
solar whale
yes
slender iron
I tried the 56 only
tidal kiln
@raven canopy thanks for info. yep. this was a retro update, so just wanted to make sure something like adabot wouldn't get unhappy.
slender iron
this afternoon is debugging time π
solar whale
I'll try the pca10056 and feather_nrf52832
raven canopy
@tidal kiln all good. i noticed a lot of example filenames don't "conform" to the new standard when adding them to the bundle. which can be confusing on some of them, just looking at the bundle. π
slender iron
I bet the 832 will fail for the same reason
my guess is that its internal flash related
tidal kiln
@idle owl are we all done with MPL?
solar whale
reverted dongle to 20181105 build -- no harm - no foul π
BTW module list on 20181105 was - ```Press any key to enter the REPL. Use CTRL-D to reload.
Adafruit CircuitPython 4.0.0-alpha.2-135-gd08747d37-dirty on 2018-11-07; PCA10059 nRF52840 Dong$
help('modules')
main busio microcontroller supervisor
analogio collections micropython sys
array digitalio neopixel_write test
binascii framebuf os time
bitbangio gamepad pulseio uio
bleio gc random usb_hid
board hashlib storage
builtins math struct
Plus any modules on the filesystem
@tidal kiln it's not on PyPI yet.
slender iron
hrm
tidal kiln
but it's all set to go, right?
idle owl
Oh. Yeah. I forgot to add it to the bundle though.
I wasn't sure what you meant by "set to go" I guess
tidal kiln
so_many_steps.gif
solar whale
@slender iron on the PCA10056 -- it boots and the module list looks complete ```Press any key to enter the REPL. Use CTRL-D to reload.
Adafruit CircuitPython 4.0.0-alpha.2-141-g97bc95183 on 2018-11-09; PCA10056 nRF52840-DK with nRF52840
help('modules')
main busio microcontroller supervisor
analogio collections micropython sys
array digitalio neopixel_write test
binascii framebuf os time
bitbangio gamepad pulseio uio
bleio gc random usb_hid
board hashlib storage
builtins math struct
Plus any modules on the filesystem
Monday is my library day, is what's essentially going to happen.
slender iron
ah, so its an m4 thing
thats good to know
I found an optimization while digging π
tidal kiln
@idle owl sounds like you'll take care of it? i'm gonna check it off
idle owl
If you want to add it to the bundle, go for it, but I'll deal with it on Monday otherwise.
but yes in terms of checking it off, go for it.
tidal kiln
meh. i'll pr a bundle add. i need to get back on that horse anyway.
idle owl
ok, right on π
solar whale
cool -- I can still mount an SDCard on the pca10056 ...
@slender iron yup feather_nrf52832 does not boot
slender iron
kk, will look at that next
solar whale
@slender iron do you want me to post issues or are these comments enough for now?
slender iron
for modules don't bother but the nrf with flash would be good
I may not get to it until monday
solar whale
OK - so for the pca10059 and feather_nrf52832 -- one issue or 2?
marble hornet
Can i partition the flash ? and then mount the second partition to just the atsamd?
manic glacierBOT
raven canopy
PSA: QuickLink to Particle Mesh DFU Instructions - https://docs.particle.io/tutorials/device-os/led/photon/#dfu-mode-device-firmware-upgrade-
haven't read it through yet...but i'm off for 3 days so. β¨
marble hornet
i'm reading from this page:https://circuitpython.readthedocs.io/en/3.x/shared-bindings/storage/__init__.html?highlight=storage
solar whale
@raven canopy thanks - still waiting for mine -- HAve good time away!!
marble hornet
it looks so close
raven canopy
we'll see how borked i can get it. π
slender iron
@solar whale I think the modules thing is because I have a smaller buffer on the samds
solar whale
OK - so were things like pulseio removed?
slender iron
oh no, its just outputting info too fast and overwriting
solar whale
ah -- that's better -- I'll check in a minut -- restoring 52832
solar whale
OK -- import pulseio works
slender iron
goes to find his 832
digital slate
Hi everyone
I was telling to @slender iron earlier today trying to get a device up and running with CP
Figured out I had a bad board on the first one, that was a relief
Now I think I'm at the point of uploading the compiled UF2 bootloader code via a JLink and I'm doubting how far I am
I am able to connect to the device via JLink, stops at "Waiting for GDB connection..." , ostensibly because there is no program file?
exotic pumice
I got my hal compiling π
digital slate
I think my next step is flashing the uf2 firmware, but was hoping for a sanity check
exotic pumice
hi
digital slate
Hi Saj
exotic pumice
you don't upload uf2 with jlink, you drag and drop onto the usb mass storage device
slender iron
@digital slate you'll need to run gdb to connect the jlink software
exotic pumice
oh nvm I misunderstood
digital slate
Ah, so second window and then load up with the program
pastel panther
@digital slate You can also use JLinkExe's loadbin bootloader.bin, 0
maybe the offset and filename is reversed, I can never remember
idle owl
@slender iron What is this? https://github.com/adafruit/Adafruit_CircuitPython_CPython
pastel panther
It's good to know how to use the jlink gdb server and gdb but for just loading .bins, I find it easier to just use JLinkExe (I think it's called JLink commander?)
slender iron
@digital slate right, I use a second tab
idle owl
@digital slate Hey! π
slender iron
@idle owl its for libraries that match cpython
idle owl
@slender iron Hmm. Ok. Thanks.
digital slate
Hi @kattni!
I guess a dumb quesiton, but I'm just loading to address zero since it's a bootloader, right?
pastel panther
yup
pastel panther
can you get your blablaBOOT?
pastel panther
yea
digital slate
yep, didn't change anything so mine says feather
digital slate
hey, it's Friday, why not a little celebratory dancing??
pastel panther
I was thinking more cabbage patch than can-can but I'll take what I can get
idle owl
@slender iron @tulip sleet The 20181109 release of the TrellisM4 build appears to have issues. I'm running the NeoPixel toggle demo, which runs fine on 20181105 (and I just got back to that build and it still works). When you put the newest version on the Trellis M4, all the LEDs turn white, and the status DotStar turns purple, and then you can't even reset the board to get back into the bootloader. The only way to get out of this state is to unplug/replug/REALLYFAST hit reset twice to get into the bootloader, and then load the older version. That took two tries as well to get it to work.
JP found it first when I had him update his board to run the demo for himself. We thought it was code, but I updated mine and the same results.
(I have to run, I'll explain it better later if you need me to, but I have to head out for the evening.)
slender iron
@idle owl when you are back please file an issue with the full filenames of what you are using. thanks!
exotic pumice
what's going on with the spi flash on metro m4? SCK and CS are connected to sercom4 and MOSI and MISO are connected to sercom0
at least that's what the schematic says
@slender iron any ideas?
slender iron
its connected to the QSPI peripheral, not a sercom
exotic pumice
oh, that's a thing?
slender iron
ya
exotic pumice
k well I guess I'll just disable that feature for now
don't feel like writing a hal around the qspi peripheral
solar whale
@slender iron @idle owl reproduced the trellis m4 issue - current master does some very odd things with the neopixels and eventaully drops USB connection.
slender iron
β¬ is a fix for you @solar whale
solar whale
I found that I could get back to TRELLISBOOT but when I loaded older image, I had to power cycle to have itcome oute of bootloader.
@slender iron for nrf52832 and dongle or just for modules list?
I'll pull the PR after travis finishes.
exotic pumice
my blinky doesn't work
d13 is connected to a led still, right?
otherwise it wouldn't be an arduino
the one marked L?
slender iron
ya
exotic pumice
oh it turns on, but only when I press reset
Interesting
it's supposed to do it in a loop forever
is the reset button lighting L just part of the bootloader?
slender iron
ya
exotic pumice
hmm, well I've got some debugging to do
solar whale
@slender iron hmmm -- metro_m4_express does not boot with pr_1325
slender iron
hrm, thats what I was testing on
solar whale
I did git fetch origin pull/1325/head:pr_1325 then built merto_m4_express -- after falsh -- Blue neopixel -- no CIRCUITPY or /dev/ttyACM0
reverted to current master OK
solar whale
ah -- DEBUG hides things
exotic pumice
I think I'm probably setting up the clocks wrong
solar whale
from past (long past) esperinces with DEBUG hiding issues it was often an unitintialzed variable.
slender iron
@solar whale pushed a fix
just added a check to see if usb is connected before we wait for everything to be sent
slender iron
yup, me too. using a tc to output a pwm you think you know can help get it tuned
@solar whale do you know how to swap bootloaders on the '832 feather? I have the mynewt version
solar whale
I build from the bootloader_nRF52 repo then flash via J-link - then in CP repo use jlink to make BORAD=feather_nrf52832 sd to get proper SD
slender iron
kk
solar whale
finally load vi make BOARD=feather_nrf52832 dfu-gen dfu-flash SERIAL=/dev/ttyUSB0 -- or just use Jlink make BOARD... flash
woohoo! ```Press any key to enter the REPL. Use CTRL-D to reload.
Adafruit CircuitPython 4.0.0-alpha.2-143-gd012fd155 on 2018-11-09; Adafruit Metro M4 Express with samd51j19
help('modules')
main busio micropython storage
_os collections neopixel_write struct
_time digitalio network supervisor
analogio errno os sys
array gamepad pulseio time
audiobusio gc random uerrno
audioio i2cslave rotaryio uio
bitbangio io rtc ure
board math samd usb_hid
builtins microcontroller socket wiznet
Plus any modules on the filesystem
yay!
solar whale
@slender iron I ran into a problem in the past - if I flashed sd from the bootlaoader repo, it used the wrong version as default -- CP expects s132 for the 52832
slender iron
I used this bin: https://github.com/adafruit/Adafruit_nRF52_Arduino_Bootloader/blob/master/bin/feather52832/6.0.0/feather52832_bootloader_6.0.0_s132.bin
solar whale
looks good
solar whale
ah -- I think CP still uses SOFTDEV_VERSION 5.0.0
see the nrf/boards folder
I think if you do a make BOARD=feather_nrf52832 sd in CP repo it will be OK
slender iron
as its connected over jlink or nrfutil?
solar whale
jlink
solar whale
the sd's and SOFTDEV's are a bit confusing....
slender iron
ya, I don't know why they give options. we should just always have one we support
solar whale
in the bootloader_nRF52 I think it is set up for Arduino -- the CP repo does have the defaults for us.
I think @gentle bronze understands it π
slender iron
ya, that was easy to update at least
covert oxide
Hello, so if I want a board to support double-precision floats, all I have to do is modify the mpconfigport.h , and st MICROPY_FLOAT_IMPL to MICROPY_FLOAT_IMPL_DOUBLE? and perhaps have enough space? is this correct?
slender iron
@solar whale gotta run. will let you know when I have a fix
solar whale
Thanks - no rush here!
timber mango
@exotic pumice Trellis M4 first Adafruit board I've seen (SAMD) that does not have D13 LED (at all).
exotic pumice
interesting
timber mango
My USB code ran first try, on Trellis M4 - I had vague plans to blink my own LED as a test for basic ops but was unneeded.
exotic pumice
does it only have neopixels?
timber mango
Has DotStar singleton, too (on the reverse side from the keyboard matrix)
May be the only Adafruit target with both DotStar and NeoPixel drivers required for full use of the board.
exotic pumice
yeah, that's tricky
timber mango
I've bit-banged DotStar APA-102 (Pimoroni Blinkt)
I still don't know how to send to the rest of the NeoPixel array -- the code I lifted only drives the first pixel.
exotic pumice
yeah but to not have a simple led for debug is pretty crummy
with neopixel data, you just send 3 bytes per pixel
so if you want to light pixel 2 send 6 bytes
timber mango
Well Trellis M4 is much like an ItsyBitsy M4 so get that going first.
I'll try that. APA102 is kind of like that as well.
exotic pumice
I'm working on a neopixel driver as well
timber mango
The seesaw code for neopixel is a lift of the arduino library - I got it from seesaw.
that nop stuff ;)
timber mango
haha well if you read the original driver comments it seems recommended.
exotic pumice
yeah, my goal is to create a driver that can control neopixels with any mcu, so it doesn't fit my vision
any mcu with dma and spi that is
timber mango
oh you're the person who wants to use hardware SPI drivers. ;)
I definitely recommend a thorough read through of the Arduino lib for NeoPixel from Adafruit.
exotic pumice
yeah I've given it a good look
timber mango
nop nop nop nop nop
I'm sure there's a Star Trek TOS episode in that. ;)
(right up there with the horta)
bold plaza
Happy Friday all! Still doing my googling, so far coming up empty.. anyone happen to have a link to a Circuit/Micro python package that supports the OLED SSD1331 (color) by chance? https://learn.adafruit.com/096-mini-color-oled/overview for reference
pastel panther
exotic pumice
My clocks are so broken I can't even get a nopping blinky to work
it probably doesn't help that a lot of my clock code is trying to achieve 48MHz
obsidian compass
Do we need to download espriff/xtensa toolchain to build circuitpython for ESP targets?
I did this already "make -C mpy-cross"
latent nimbus
Good day all,
I have a question about NeoPixels on CircuitPython on Raspberry PI. Why must the code run under sudo?
solar whale
@latent nimbus not sure of the details but it was confirmed in this discussion https://github.com/adafruit/Adafruit_Blinka/pull/32
raven canopy
@obsidian compass MicroPython and CircuitPython use a 3rd party SDK (includes espressif's toolchain) called esp-open-sdk. build steps for the esp8266 port are here: https://github.com/adafruit/circuitpython/tree/master/ports/esp8266#build-instructions
obsidian compass
Thanks for sharing. @raven canopy π
latent nimbus
@solar whale Thanks...
raven canopy
@solar whale the more i read up on the Particle Mesh and DeviceOS, its seems that to utilize it with all it offers we'll have to make a new port and somehow glue in the Particle Cloud stuff. otherwise its just a Feather nRF52. π΅
raven canopy
they've set it up in an interesting way. i think i understand why, and that was not a derogatory judgement. at least the firmware is up on GH: https://github.com/particle-iot/firmware/tree/mesh-develop
pastel panther
Either the crystal or two caps next to each other, across the same lines that were shorted with each other
I forgot that I had @timber mango 's code loaded so after I removed the crystal and its caps and then didn't see a CIRCUITPY drive, I kept going and adjusted the caps. Then when I was hopelessly poking around I noticed a /dev/tty.usbmodemxxxx!
I was also preparing to desolder the samd and was warming it up when I fixed the caps, so theoretically I could have reflowed an iffy connection there as well
Now I just need to see about getting uf2 and CP back on
slender iron
@pastel panther yay! we generally don't use an external crystal which simplifies things
@exotic pumice it runs at 48 by default I think
pastel panther
It's not working again now but I'm guessing it's a software issue cause I've been loading bins willy nilly
slender iron
π
timber mango
Yeah it definitely runs at 48 MHz by default, @slender iron
But there's also an 8 MHz thing going on (probably for M0, don't remember).
Depends on the code base, perhaps. ;)
needs an oscope that can go higher than 25 MHz ;)
pastel panther
(β―Β°β‘Β°οΌβ―οΈ΅ β»ββ»
timber mango
The scope I already have displays a 48 MHz signal kinda/sorta.
sidd: 48 MHz is probably difficult to get away from more than hard to achieve in the first place.
(USB wants it for one thing)
I think the trick is to get above 48 MHz (MCU clock) while keeping USB at 48 MHz.
Think I saw this in circuitpython tree (maybe not actively engaged):
https://github.com/wa1tnr/ainsu-CamelForth-d51-usb/blob/master/boards/metro_m4_usb/driver_init.c#L110
pastel panther
welp, I'm back to things solidly not working again π
I think I'm just going to write this board off and try to assemble a new one
timber mango
$ pwd | cut -b17-99
Adafruit/circuitpython
$ ag 25% | ag atmel
ports/atmel-samd/asf4/samd21/driver_init.c:386:/* The USB module requires a GCLK_USB of 48 MHz ~ 0.25% clock
ports/atmel-samd/asf4/samd21/driver_init.c:389:#warning USB clock should be 48MHz ~ 0.25% clock, check your configuration!
ports/atmel-samd/asf4/samd51/driver_init.c:422:/* The USB module requires a GCLK_USB of 48 MHz ~ 0.25% clock
ports/atmel-samd/asf4/samd51/driver_init.c:425:#warning USB clock should be 48MHz ~ 0.25% clock, check your configuration!
I don't know if that asf4 tree is really in the compiled image or not.
pastel panther
or maybe just write off hardware entirely and go back to fishing
timber mango
poisson!
pastel panther
les petit poisson?
timber mango
ladyada_ says to get used to setbacks. ;)
She's said that two or three times in various videos (in differently-worded utterances)
pastel panther
I'm not not used to it, I'm just not good at channeling my inner duck
timber mango
get one of those giant BFC capacitor ducks ;)
velvet badger
I liked Mr. LadyAda's "Software works until it doesn't, hardware doesn't work until it does."
It's a reoccurring theme in my projects.
Or maybe it's the reverse. scratches head
digital slate
found a broken page
idle owl
Guide page?
digital slate
going from here
the next button to the pycharm page is broken: https://learn.adafruit.com/adafruit-feather-m4-express-atsamd51/pycharm-and-circuitpython
idle owl
Oh right.
That page has been removed because it's not valid anymore. It shouldn't be trying to use it anymore, so that's odd.
Ah I see why. Fixing it now. @digital slate Thanks for catching that!
digital slate
no prob!
idle owl
Should be good now
idle owl
Ok I've repro'd the TrellisM4 NeoPixel issue multiple ways. Filing an issue in a bit here.
lime trellis
Is it possible for CP to use spi flash as a memory partition?
raven canopy
Hamtonio lives! π
timber mango
@pastel panther awesome!
pastel panther
when in doubt, throw it in the trash and start over
timber mango
I went for a few months, maybe, on the UART before tackling USB.
pastel panther
I'm really excited to dig into your usb code. Thanks much for sharing
timber mango
You're welcome!
rare zephyr
Question does circuitpython have any moduals to operate as logic gates or like PLC?
timber mango
It has no ringbuffer so a huge kludge is used -- long delays.
The USB stuff seems to complete in the background on its own; the delays make sure no new use of USB occurs while that is happening.
pastel panther
@rare zephyr not that I know of
Though I don't know precisely what you have in mind
timber mango
Well you can do logic operations and control GPIO to correspond the what the program does with your inputs.
pastel panther
If you want to control relays programmatically, it can do that but it's all done by writing python
timber mango
(that's true of any and all microcontrollers, generally)
pastel panther
what nis said is also true
rare zephyr
i am looking to be able to detect a blinking light with a photo sensor and if it is blinking at 1hz turn on an out put pin but if it is on solid turn off the output pin
pastel panther
depending on the accuracy you want, that should be possible
timber mango
The problem with using any microcontroller's input pins are very different from the problem of using its outputs. ;)
Generally you're looking at scanning, or interrupts (or both) or busses (serial I/O).
Whereas a standalone logic gate just 'scans' though we don't call it that.
manic glacierBOT
errant grail
@pastel panther Excellent!
pastel panther
@errant grail I haven't had the huevos to put the boost converter together on this one yet
timber mango
@pastel panther make install make clean and other make targets exist in most of my projects.
They're all in the final 100 lines or so of a given Makefile
pastel panther
good to know; does install use bosac?