Whee I broke my timing.
#circuitpython-dev
tulip sleet
I even get an email: Subject: [adafruit/circuitpython] Run failed: Build CI - 8.1.0 (b1a28bc)
but only because I initiated those runs
idle owl
Oh, come on. π€¦π»ββοΈ
I removed the thing that reset the time check for that block somehow.
crimson ferry
2038, right, that's the end of time (I think) for CircuitPython```py
sys.maxsize
2147483647
time.localtime(2147483647)
struct_time(tm_year=2038, tm_mon=1, tm_mday=19, tm_hour=3, tm_min=14, tm_sec=7, tm_wday=1, tm_yday=19, tm_isdst=-1)
brazen hatch
2038, right, that's the end of time (I think) for CircuitPython```py
>>> sys.max...
pretty sure I had tested it and it worked passed it.
idle owl
@tulip sleet Turns out 6 pings per second is too much for the ESPIDF. π
@crimson ferry OK, so I made this work, but kept my original code intact, like a smart person for once. Are you saying this is completely unnecessary, and I should simply use time.time() in place of monotonic()?
That's completely fine, but I wanted to clarify before I revert all of this, llol.
crimson ferry
should be fine for rough-one-second intervals
idle owl
should be fine for rough-one-second intervals
Ok. Thank you.
crimson ferry
no floats, no rollover, no extra dependences
jaunty juniper
pretty sure I had tested it and it worked passed it.
ints have no size limit, but time functions are limited:
>>> time.localtime(2147483647+1)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
OverflowError: overflow converting long int to machine word
FWIW, it was keeping in time with the RTC intervals in the time I let it run.
e.g. I had it printing time.time() and monotonic() at the same time as the code block print, and it was exactly 1 second apart.
crimson ferry
the code might have gotten sync'd up to a second boundary
(or drift due to float precision - but not likely if you've been resetting it)
oh, n/m, I think I misunderstood
idle owl
I'm saying the time.time() interval was matchy with the RTC 1 second interval.
Which is a good thing, I believe.
crimson ferry
yeah, I believe everything comes from the same clock internally?
idle owl
Fair enough.
This was a much simpler change. Thank you. π
But I got RTC working! Saving that code. Because that was a whole ordeal. And I can crib it later.
My ping times here are showing 0.0ms maybe 75% of the time. Is that seriously right?
Am I having a good internet day? Or is CircuitPython lying or something, heh...
crimson ferry
RTC is nice, on espressif even more nice since it will keep real clock time across reloads AND soft resets, making nice timestamps once you set the RTC
idle owl
RTC is nice, on espressif even more nice since it will keep real clock time acro...
Good to know.
crimson ferry
ints have no size limit, but time functions are limited:
```py
>>> time.localtim...
CircuitPython 53.0.0 in 2038 will only run on 64-bit processors :p
brazen hatch
ints have no size limit, but time functions are limited:
```py
>>> time.localtim...
wait, how does rtc handle it then?
rtc.RTC().datetime = time.struct_time(...)
I can set it just fine with my date implementation
It just works.
jaunty juniper
interesting
brazen hatch
CircuitPython 53.0.0 in 2038 will only run on 64-bit processors :p
Looking forward to when Circuitpython competes in the UEFI market.
manic glacierBOT
idle owl
If anyone is up for a review: https://github.com/adafruit/Adafruit_Learning_System_Guides/pull/2512
jaunty juniper
note that the error I showed would be a limitation of the implementation of time.localtime(), not a limitation of struct_time itself
brazen hatch
note that the error I showed would be a limitation of the implementation of time...
sounds like an easy fix
i will try switching it
midnight ember
ints have no size limit, but time functions are limited:
```py
>>> time.localtim...
I ran into that same overflow error working with the gps time. never did figure out how to fix it. it would happen randomly. :/
jaunty juniper
I ran into that same overflow error working with the gps time. never did figure ...
erroneous character in a UART communication leading to an absurd value or something maybe ?
brazen hatch
Wait, doesnt CP do parity bit?
jaunty juniper
does the GPS do it ?
empty salmon
I like the AVR chips for their price, low power capability, and availability.
manic glacierBOT
midnight ember
As far as i could tell it seemed like gps data was trying to merge with the time struct. I donβt know how else it would overrun the int size.
I didnβt know about error handling then. Might be able to revisit it.
brazen hatch
It's ready for review. How was Camera not a feature?
Have there really been that few CP boards with a camera?
midnight ember
ESP used to be the only one powerful enough to do a camera of any kind, maybe a few others. yeah there aren't many.
now with supercharged boards like iMX we might see more higher power projects.
brazen hatch
now with supercharged boards like iMX we might see more higher power projects.
higher power bloat π
I am the king of bloat & I approb.
manic glacierBOT
brazen hatch
Ok, guys, I want you to hear me out.
I feel a speed difference between .bin and .uf2.
Like, actual difference in speed of the board during operation.
the uf2 feels faster???
I seriously need to make boardbench.
I have no way to prove it right now.
thorny jay
It's ready for review. How was `Camera` not a feature?
I am pretty sure Camera was already suggested by me.
Either in a meeting, or as an issue.
And the answer could have been: "We need to limit the number of feature." + "Only a few board have Camera, and people looking for that will get the information from the learn guide... or the feature matrix."
But that is from memory, I could be wrong.
brazen hatch
Well, it's in the PR. I guess I will find out soon enough.
And the argument that there aren't that many won't stand for long.
m5stack has like 14 boards with cameras, all of which should eventually get a port
https://shop.m5stack.com/collections/m5-cameras
crimson ferry
not sure where to ask for a semi-official opinion... Does Adafruit have a preference for which pins on a Feather (RP2040) should be Wire1 in Arduino? https://github.com/earlephilhower/arduino-pico/issues/1465 D6/D9 and D112/D13 are used by a lot of featherwings, RX and TX probably not, D24/D25 are SWD
onyx hinge
thanks Dan
tulip sleet
not sure where to ask for a semi-official opinion... Does Adafruit have a prefer...
cater is following up on this, thanks
tidal kiln
oh. yep. there's discussion scattered in several threads at this point. but TLDR = going to make breaking change so that SCL/SDA are Wire. here's the PR that will do that:
https://github.com/earlephilhower/arduino-pico/pull/1468
in terms of "preference" - most definitely that SCL/SDA be Wire.
crimson ferry
@tidal kiln thanks for getting involved, it was quite confusing
tidal kiln
<@337989520962093057> thanks for getting involved, it was quite confusing
yep. very much so.
proven garnet
Just a heads up to everyone I'll be patching the pylint and jQuery (RTD) issues this evening, so those issues should be resolved within the next few days. Also to merge/rebase if needed.
Though the latter point shouldn't apply to most if anyone I think.
crimson ferry
@graceful rain dev stuff better in here... you can put it in the make command line, but you'll need to turn off some modules to allow the debug build to fit on flash, see #circuitpython-dev message
backtrace decode utility is at ports/espressif/tools/decode_backtrace
graceful rain
thanks @crimson ferry wow this is really getting into the weeds. Where does this decode_backtrace go in my own test program? I'm not too sure what I'm looking at or how to use it, haha. "Enter the backtrace line at the "? " prompt." - what does that mean?
crimson ferry
first, you have to hook up your host to the debug serial pins and monitor that (19 & 20 I think on the S2?), then when you get a safemode / hard fault, there will be a list of addresses... those get pasted into decode_backtrace and out will come a stack of the C code line numbers
graceful rain
Ok, thanks. So that decode_backtrace code is running on the host?
graceful rain
Sorry one more question. How does that code know what serial port to listen to. I don't see it defined anywhere
brazen hatch
Your question is vague.
In your os? In terms of board pins?
graceful rain
@crimson ferry said to hook up the debug serial pins and monitor that. But I don't see anything here (https://github.com/adafruit/circuitpython/blob/main/ports/espressif/tools/decode_backtrace.py) about monitoring serial pins or defining the serial port.
Yes, in my OS (Linux)
slender iron
the python script doesn't connect to it
just use your normal serial terminal
and then copy the backtrace line over from there
crimson ferry
does the backtrace go to the regular serial console? I thought it went to the debug console
I always hook up the second USB to get debug messages and safemode backtraces
slender iron
does the backtrace go to the regular serial console? I thought it went to the de...
it goes out the uart because it is done by the IDF
orchid basinBOT
Please don't try and circumvent the checks by adding to the features list without previous agreement. The reason we don't have it on the list is because there are so few boards with this feature and we're trying to keep the features filters from being too long:
graceful rain
@crimson ferry got a debug build up, and observed debug serial, but it didn't really show anything aside from the usual:
Build:Oct 25 2019
rst:0x3 (RTC_SW_SYS_RST),boot:0x8 (SPI_FAST_FLASH_BOOT)
Saved PC:0x4002b865
SPIWP:0xee
mode:DIO, clock div:1
load:0x3ffe6100,len:0x1688
load:0x4004c000,len:0x1048
load:0x40050000,len:0x2f04
entry 0x4004c350
I (24) boot: ESP-IDF v4.4-dev-3608-gbbe2a1bf34 2nd stage bootloader
I (24) boot: compile time 23:14:23
I (24) boot: chip revision: 0
I (28) qio_mode: Enabling default flash chip QIO
I (33) boot.esp32s2: SPI Speed : 80MHz
I (38) boot.esp32s2: SPI Mode : QIO
I (43) boot.esp32s2: SPI Flash Size : 4MB
I (47) boot: Enabling RNG early entropy source...
I (53) boot: Partition Table:
I (56) boot: ## Label Usage Type ST Offset Length
I (64) boot: 0 nvs WiFi data 01 02 00009000 00005000
I (71) boot: 1 otadata OTA data 01 00 0000e000 00002000
I (79) boot: 2 ota_0 OTA app 00 10 00010000 00160000
I (86) boot: 3 ota_1 OTA app 00 11 00170000 00160000
I (93) boot: 4 uf2 factory app 00 00 002d0000 00040000
So it hasnβt gone into safe mode yet?
When it does, the crash dump will print there with the traceback addresses
orchid basinBOT
tulip sleet
@onyx hinge @slender iron I am going to make an 8.2.0-beta.0. Release notes are all done. I don't see any imminent PR's. OK by you?
onyx hinge
@tulip sleet I have nothing pending for synthio that affects behavior, just an added set of tests
tulip sleet
sounds good; the main point of this is to refresh synthio and get some new boards in
slender iron
<@426200547024961539> <@252717193496756235> I am going to make an 8.2.0-beta.0. ...
All good to me
proven garnet
@idle owl heads up that adabot failed the bundle cron last night. I'll dig into it, but letting it fail again might help. I'll have it fixed this week though.
idle owl
<@330227457296957440> heads up that adabot failed the bundle cron last night. ...
Ok.
Thanks.
proven garnet
π«‘ its prooooobably NAU7802
idle owl
Agreed
proven garnet
But the error text was... lacking in helpful information. I'll likely need to poke it locally
Actually, did anyone turn on branch protection for the Adafruit Bundle repo recently? I re-ran it and it's giving me the same issue.
I think Adabot is supposed to be able to write to the main branch but can't now for some reason. At least that's what the error suggests.
idle owl
OK, I'll check.
proven garnet
Oh that might be the old action again, but that's what the error was the first time at least
We'll see what re-running does
Yup, same thing.
idle owl
I am not seeing anything indicating we changed something.
proven garnet
hmmmmmmmm, it definitely reads like a permissions issue with actually pushing the changes.
STDERR:
remote: Resolving deltas: 0% (0/3)
remote: Resolving deltas: 33% (1/3)
remote: Resolving deltas: 66% (2/3)
remote: Resolving deltas: 100% (3/3)
remote: Resolving deltas: 100% (3/3), completed with 3 local objects.
remote: error: GH006: Protected branch update failed for refs/heads/main.
remote: error: At least 1 approving review is required by reviewers with write access.
To https://github.com/adafruit/Adafruit_CircuitPython_Bundle.git
! [remote rejected] main -> main (protected branch hook declined)
error: failed to push some refs to 'https://github.com/adafruit/Adafruit_CircuitPython_Bundle.git'
Do me a favor, and take a look? And do a google search on what setting would block that? Because I did both, and not seeing anything or getting a useful search result.
manic glacierBOT
proven garnet
Do me a favor, and take a look? And do a google search on what setting would blo...
Sounds good. It might be part of a larger question of what repos should have that protection. Doing patches, I noticed a lot had protections against pushing directly to main that I was curious why (neopixel I understand, but there seemed to be a bunch of random other less popular libraries that had it)
idle owl
Yeah I was seeing that.
We need to fix that or figure out excluding Adabot only from that. Except on neopixel. π
proven garnet
I can look into that.
idle owl
Thanks. But let's figure this out first.
Wait.
It does say "require a PR before merging".
idle owl
But you can create roles that are excluded. Thing is, who is Adabot running as? She's not an actual user I think? So how do we create the exclusion role since I think it runs under multiple users?
proven garnet
It does say "require a PR before merging".
Bingo!
proven garnet
But you can create roles that are excluded. Thing is, who is Adabot running as? ...
mmmmmmmm, good question
proven garnet
This should be adabot: https://github.com/adafruit-adabot
So giving her an exception should be sufficient for CI purposes
slender iron
adabot should be a separate account
tulip sleet
It does say "require a PR before merging".
I recently added branch protection to a bunch of libraries, and maybe the bundle repo too.
I can turn that off immediately. I did it because GitHub started warning about unprotected branches. In regular libraries this makes sense: a library update should have a PR. We can add exclusions.
We can give adabot (adafruit-adabot?) this bypass permission:
proven garnet
I can turn that off immediately. I did it because GitHub started warning about u...
I think that should work. Because then at least someone could log onto adabot, get a temp token for patches, and run with that.
I like making sure no one can write to main, but that's also what will happen a lot haha
tulip sleet
I'm sorry I caused trouble here; I didn't realize there was direct pushing to main on the Bundle.
proven garnet
No worries! Glad it's not a chase down the rabbit hole if anything π
tulip sleet
so what privileges does adafruit-adabot need on the Bundle repo? I think Admin.
proven garnet
Is that sweeping permissions? The ones I can think of are anything relating to contents, Actions, and workflows.
proven garnet
That would also be great, and might be futureproof for adding more people
If that ever happens
tulip sleet
uhh, can't do it without being GitHub Enterprise, which we are not
proven garnet
Excluding adabot would help at least
Not sure the difference between admin and maintain though in terms of permissions
tulip sleet
admin can delete the repo, etc. basically superuser on the repo
maintain is not good enough to bypass protections.
So I can make adabot be admin, or take away the rule
proven garnet
@idle owl, what are your thoughts when you get a chance?
Even if you take away the rule, adabot couldn't delete, right?
tulip sleet
ok, I did this:
- Make adafruit-adabot have Maintain.
- Turn off all the branch protection stuff (PR required, etc., except this:
I turned this on. I think this will allow PR merges from Librarians, and allow direct pushes from adabot. Is there an easy way to test the adabot role now?
I could re-run the failed job from last night?
I'm not sure that Librarians can merge PR's to Bundle now, with that set.
So worst case, we just delete the protection rule, if this doesn't work.
proven garnet
I turned this on. I think this will allow PR merges from Librarians, and allow ...
I can try the bundle CI again and let you know!
tulip sleet
actually, for 2. above, I added Librarians to who can push, so that should be fine now.
tulip sleet
I can try the bundle CI again and let you know!
I just clicked save, so if it failed a few seconds ago, try agian
tulip sleet
yay! I now understand this better. One more q -- sometimes we do massive changes across all libraries. I thought these were sometimes done without PR's. Is that (still) true? Are the branch protection rules getting in the way of that? Your latest change was all PR's.
proven garnet
yay! I now understand this better. One more q -- sometimes we do massive changes...
Adabot can do many of them, but if they can't be patched (via git format-patch) then they'll come from a human account
All those PRs I did were pretty much exclusively due to branch protection. I might PR changes that do require a second look like pylint updates but yesterdays weren't in that category.
tulip sleet
but the patch is a direct push to main, is that right? On libraries without branch protection rules that would be no problem, but since I've been adding those rules, then that won't work any more. So I should allow adafruit-adabot to push directly in the branch protection rules, and then it will work agian.
i don't know if there were no branch protection rules before, but maybe that was the case.
proven garnet
but the patch is a direct push to main, is that right? On libraries without bran...
Yeah the patch is a direct push to main, I just wanted to highlight who was pushing.
So yeah, allowing adafruit-adabot will fix those and prevent any pushes from humans, but that also means any pushes that are hand-tuned or manual pushes (maybe the patch could be applied like yesterday's to many libraries), it now means all those will need to be PRs now. Before I used to be able to just push directly to main.
proven garnet
I think I can sidestep it if I'm provided a PAT and authenticate as adabot, so that's one solution if we want to keep those protections but temporarily allow manual fixes like yesterday's to avoid generating a ton of PRs. Or not, I'm open to solutions of any kind.
The other case is I have scripts for doing patches that are patched by parsing repository contents using Python as opposed to git format-patch, so those also come from a human and not adabot. So same things as above apply there. An example of such a patch would be change all SPDX authors to FooBar - git format-patch would fail so I use parse in Python to iterate through all the libraries and do it, checking, writing, committing, and pushing.
idle owl
@proven garnet @tulip sleet Thanks for the fix! If we need to do something more broadly with the libraries, please loop me in.
manic glacierBOT
manic glacierBOT
cerulean fractal
Hi! I have a question: why aren't there any ports for MCU modules? The kind you solder onto a carrier board. What are the challenges for supporting these?
Or is it the question of target audience?
lone sandalBOT
stuck elbow
Hi! I have a question: why aren't there any ports for MCU modules? The kind you ...
there is no way of telling what pin will be used for what, one big point of a board definition is that it defines pin names
if you build a board using a module, you can add a board definition for it
cerulean fractal
there is no way of telling what pin will be used for what, one big point of a bo...
I feel we can just copy microcontroller pin names, it's already being done for some breadboardable modules anyway
jaunty juniper
Hi! I have a question: why aren't there any ports for MCU modules? The kind you ...
because nobody has made one yet. There are some board definitions for Espressif dev boards that bring out all the pins, which can generally be used for the compatible module.
Then again, pure modules are not super practical, considering how CP generally relies on a USB port or at least a dedicated serial (with a serial-to-usb chip). You wouldn't really be able to use a generic build without specifying at least one of those
cerulean fractal
Yeah, makes sense... There's usually only one USB interface, that can probably be assumed as default
manic glacierBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
lethal grove
Hey folks, anyone want to help with some under-the-hood circuit python library extension stuff?
I'm trying to add a library to CircuitPython (the library in question is from Neuton, a TinyML platform - https://lab.neuton.ai )
My team and I used to have this library compiling, but Neuton went and changed how they distribute their code. Previously, they gave a file tree that was
/YourModel
/Artifacts
/neuton
/include (h files)
/lib (c files)
/neuton_generated
But they've changed it so that /lib contains .a files (https://fileinfo.com/extension/a) - and now CMake is unhappy with our old tools for binding the library into CircuitPython claiming that it can't find a rule to make Neuton.C, which now doesn't exist.
How should I direct CMake to parse a ".a" file.
Even a start on what to read up on would be helpful because it's currently bottlenecking our project and I don't know how to fix it.
Thanks in advance!
orchid basinBOT
brazen hatch
Hey folks, anyone want to help with some under-the-hood circuit python library e...
iirc .a files are assembly
if that is the case, unless it's arm/riscv code that ain't gonna work, even if you get cmake going
or was it .S It's 6am with no sleep, idk anymore
brazen hatch
I feel like wifi reconnection should be something like a periodic check.
If it has connected once successfully, it should do a bunch of attempts at once when it's lost. And then more attempts at intervals within backround_tasks
my engluish is dying, 8am after allnighter
Doing a bunch of attempts when a problem has just occured is not really helpful in most scenarios.
crimson ferry
yeah, we should probably entertain some other strategies
tough call though, at some point you may want to send it back to user code to make a decision
brazen hatch
As it stands now, it 'wants' to be seemless.
Which it really isn't.
But it's a good time to decide on api changes cuz 9.0.
crimson ferry
I seem to recall some discussion of whether web workflow auto-reconnect should trigger more aggressive always-connected approach
and then there are still differences in behavior between espressif and raspberrypi
brazen hatch
and then there are still differences in behavior between espressif and raspberry...
Yea, I really need to at some point go back to rp2 and do the different wifi security stuff for ap.
The thing is just that I am now so happy with S3 that picow just feels sad in my eyes.
crimson ferry
Adafruit CircuitPython 8.1.0-beta.2 on 2023-04-26; Adafruit QT Py ESP32-S3 no psram with ESP32S3
>>> import os, wifi
>>> wifi.radio.connect(os.getenv("WIFI_SSID"), os.getenv("WIFI_PASSWORD"))
>>> wifi.radio.connected
True
>>>
soft reboot
Auto-reload is off.
Code done running.
Press any key to enter the REPL. Use CTRL-D to reload.
Adafruit CircuitPython 8.1.0-beta.2 on 2023-04-26; Adafruit QT Py ESP32-S3 no psram with ESP32S3
>>> import wifi
>>> wifi.radio.connected
False
```vs.```py
Adafruit CircuitPython 8.1.0-rc.0-6-g6cfdbfbdd on 2023-05-21; Raspberry Pi Pico W with rp2040
>>> import os, wifi
>>> wifi.radio.connect(os.getenv("WIFI_SSID"), os.getenv("WIFI_PASSWORD"))
>>> wifi.radio.connected
True
>>>
soft reboot
Auto-reload is off.
Code done running.
Press any key to enter the REPL. Use CTRL-D to reload.
Adafruit CircuitPython 8.1.0-rc.0-6-g6cfdbfbdd on 2023-05-21; Raspberry Pi Pico W with rp2040
>>> import wifi
>>> wifi.radio.connected
True
(web workflow off in both cases... espressif disconnects on reload, raspberrypi does not)
brazen hatch
That's cuz on rp2 it never resets the wifi stuff underneath.
Not much can be done with the current code.
crimson ferry
your breakfast time, my sleep time (don't know if we had crossed paths yet, but I was in your city last summer π
brazen hatch
A rewrite would be needed.
And with the current api, that would be pointless.
brazen hatch
your breakfast time, my sleep time (don't know if we had crossed paths yet, but ...
You never know.
The world isn't too big.
brazen hatch
Discord could do a lot more to be a professional tool.
but doesnt, prolly intentionally
Business discord would annihilate slack
humble shadow
Hey, I might need some help with editing these files for a new circuitpython board i'm developing. Before I get into it, is this the right channel for these types of questions?
random junco
Hey, I might need some help with editing these files for a new circuitpython boa...
This is the right place, you don't need to ask to ask. π It is a holiday weekend in the US, so devs might not be around on a Sunday morning. Almost all of us are volunteers, so I'd recommend asking your question and when someone sees it and knows, they'll speak up
humble shadow
This is the right place, you don't need to ask to ask. π It is a holiday week...
Sounds good, thanks π
quaint monolith
I've developed a system where I control mouse movement from a Windows device connected to my Raspberry Pi via a socket connection. When I send the movement coordinates (x, y) from the windows device to the Raspberry Pi, the messages sent go so fast, they manage to make the mouse move but slip and don't have time to be properly executed.
I've tried sending the coordinates sequentially with a delay (time.sleep) between each transmission. This works, but it slows down mouse movement considerably, which is undesirable as I jerk and lose a lot of fluidity.
I also tried using an execution queue to process the coordinates sequentially, but this didn't solve the smoothness problem.
I'm now looking for advice on the best approach to solve this problem and make mouse movement smoother, without introducing additional delays or latency.
I would greatly appreciate any help or suggestions to improve the smoothness of mouse movement in my system.
Thank you in advance for your assistance!
brazen hatch
Turn mouse acceleration off from sys settings.
To make it look like human movement send many coords of smaller steps to the pc, or even make it go to (x,y) in a curve.
Disclaimer: I pulled an allnighter, may have misread the wall o' text.
The trick virtual machine software use is to register the hid device as a pen-touch tablet.
So that regardless of os, it doesn't overshoot.
Pretty sure it's not in CP as of now though.
Also #help-with-circuitpython for code.
Here is for CP development itself.
@quaint monolith
quaint monolith
Turn mouse acceleration off from sys settings.
To make it look like human moveme...
I had already done it
brazen hatch
And it still overshot?
Weird.
I haven't touched usbhid in a bit, but last I checked it worked fine.
quaint monolith
yes, my mouse receives the coordinates and moves but it drifts
quaint monolith
Weird.
I haven't touched usbhid in a bit, but last I checked it worked fine.
but I send several hundred coordinates per second
but I can't help but send so many
brazen hatch
I mean, that really isn't a great idea.
How about implementing some sort of 'fps' cap?
Like, every second I want to send n updates max.
So divide 1 by n.
Setup a time check func that reads last update time & time monotonic.
If the threshold isn't met, discard the update.
This way it can just send updates based on that n, regardless of the speed of the code.
I suggest starting with a small n like 10.
quaint monolith
infact every movement of my mouse means that I have to update the coordinates sent to my raspberry. So if I have 144 fps I have to try to send as much as possible so as not to feel the latency
brazen hatch
During development I suggest you don't be overzealous with speed.
Optimisations can be done later, and the handicaps can be removed at any time.
quaint monolith
I don't think I fully understand your proposed solution to help me. I don't know if it's the translation into my language that doesn't make me understand, but I'm having trouble.
manic glacierBOT
quaint monolith
I mean, that really isn't a great idea.
How about implementing some sort of 'fps...
in what you told me I can't Limit the number of sends per second and Send multiple coordinates of small steps
brazen hatch
I didn't say any "you can't".
What I said is that you should limit the number of 'sends' per second.
lethal grove
iirc .a files are assembly
if that is the case, unless it's arm/riscv code that ...
.a files are archive files, are a bunch of .c.o files
An A file contains a library of functions and headers that may be referenced by a C/C++ source file. It may store only a few functions or may include an entire library of functions, such as a 3D modeling engine. A files are typically created by the GNU ar utility.
manic glacierBOT
manic glacierBOT
thorny jay
yes, my mouse receives the coordinates and moves but it drifts
Did you try absolute mouse hid-descriptor?
There is a library for doing that in the community bundle.
This might ease what you try to do.
manic glacierBOT
vague thicket
Hello, is here anybody who could point me to some tutorial for writing/porting epd driver or describe me how to debug SPI data for it? I asked in help with circuitpython, but except @slender iron (thanks!) nobody respond. Display is working under Arduino, but I would like to use it with CPY. It should have Il0371 driver, and I think I have right datasheet for it and I'm also comparing some part (start sequence) to working code in Arduino. I have logic analyzer connect to my ESP32-S3 dev board so I can see what is happening at the bus and compare it to the Arduino code. I have this V1 display (https://www.waveshare.com/wiki/7.5inch_e-Paper_HAT_(B)_Manual#Introduction ), which should have this datasheet ( https://www.waveshare.com/w/upload/2/29/7.5inch_e-paper-b-Specification.pdf ) and I'm trying to change Il0373 driver from Adafruit. I also tried to search this, but without luck. Many thanks for response.
quaint monolith
Did you try absolute mouse hid-descriptor?
There is a library for doing that in ...
Let's try now
spiral storm
What's the best way for including assets (e.g. mp3 files) to be added to the filesystem on init? You can explicitly write small files like the default code.py, but that'd be messy for something as large as an mp3. Is there a path for doing this similar to how we include frozen python modules?
blissful pollen
What's the best way for including assets (e.g. mp3 files) to be added to the fil...
I believe there is a tool to create a UF2 from your entire setup. I can't recall what it is offhand as I have never used it. That may work
spiral storm
I believe there is a tool to create a UF2 from your entire setup. I can't recall...
Yeah, I was thinking that as an option. nRF Connect can show you the current memory layout and save it to a hex file. Was wondering if I could just build with the asset.
proven garnet
What's the best way for including assets (e.g. mp3 files) to be added to the fil...
I think I wrote a tool for this, but it might only be for GitHub CI....
Ah nevermind, it just creates a filesystem image file, not a UF2.
dusk mist
Hello, is here anybody who could point me to some tutorial for writing/porting e...
Since we're in the same boat here with trying to port this driver to be displayio-compatible (given that we chatted before about this and I'm working on the 2.13 display), I can help you out up to a certain point because I got mine working... with some caveats. I'm using 8.2 (though I started in 8.1 and it works there too) I got the init to work, but I couldn't get it to refresh the display the normal way.
Everything checks out according to the probe though except for the part where it uploads the framebuffer to RAM, but that's a displayio quirk for good reason. What doesn't make sense is that it won't refresh the display even when I supplied the bitpacked commands. The probe says it's coming in but it just doesn't happen. Knowing that my custom driver works with no delay included, I pretty much just packed everything into the init code and sure enough, it works. I think it's something between the init and the refresh commands that's screwing it up. Possibly the upload to RAM but I dunno. I need to investigate it, maybe even include the busy pin because I know waveshare devices kinda mess that up too.
I just realized that it only does that when I hard reset it so that's something to dig into as well. XD
dusk mist
I just realized that it only does that when I hard reset it so that's something ...
I finally got it! It's the baudrate. Turns out, the custom SPI method I was using was at the right speed at 4000000βor rather, it hits a speed wall around that point that works just fine. But on displayio, I have to crank it down to something like 10000 because it's way too fast otherwise. Like x10 too fast on 4000000 based on the timings. Now it works. Still needs to hard reset but I'm finally in. XD
lone sandalBOT
slender iron
Hey folks, anyone want to help with some under-the-hood circuit python library e...
We don't usually use cmake to build CP. .a files are already compiled so they are linked into CP . You'll want to add the .a to the linking command.
slender iron
I finally got it! It's the baudrate. Turns out, the custom SPI method I was usin...
- @vague thicket did you get it going? I'm back from the weekend.
dusk mist
+ <@340398038529081346> did you get it going? I'm back from the weekend.
I guess I ran into the same issues people had with this display showing flipped/mirrored images with displayio and eventually rediscovered (as others seem to have) that the adafruit_ssd1680 driver perfectly works fine with the 2.13 inch. I still think Waveshare's implementation clashes with displayio; currently, I can get it to refreshβnot display the images I want, but surprisingly the "code is done running" message and at any baudrate. But I don't think it matters much in this case because it's probably their init sequence, which the ssd1680 completely sidesteps.
Not a waste of time though since my main goal is to get the 7.5 inch working (I think dronecz and I have the same model) and this gives me a lot of info to go with. Maybe the ssd1680 works there too but I haven't started yet.
slender iron
so the ssd1680 sequence works ok
dusk mist
Yeah. I'm totally surprised because its init sequence is pretty simple, which makes me think these displays might have some level of ubiquity with them.
onyx hinge
<@&356864093652516868> we'll be having the weekly meeting in round about an hour. Please add your info to the notes doc if you haven't yet! https://docs.google.com/document/d/1VcPOjquw3lEHudLutRlrpuA2AQNsrTsxyYjwu2nnvZY/edit?usp=sharing
@lone axle are you good to run the meeting today even though it's tuesday?
lone axle
<@382939733107408897> are you good to run the meeting today even though it's tue...
Yep, I ended up taking today off from my other job. Thank you
vague thicket
+ <@340398038529081346> did you get it going? I'm back from the weekend.
I did not π¦ I tried to change code from @dusk mist , but without luck. Strange for me is that, that display does completely nothing in CPY, but it works in same setup in Arduino. IΒ΄m also using this code (https://github.com/lemariva/uPyEINK/blob/master/frozen_modules/e7in5.py ) as reference as far as I can tell it is pretty similar to Arduino implementation (https://github.com/ZinggJM/GxEPD2/blob/master/src/epd3c/GxEPD2_750c.cpp) .
I changed MPY driver to this (comments may not be right).
I watched today stream from @lone axle where he was fighting 7 color EPD and it did not look that hard, even when he was bit lost as IΒ΄m right now.
lone axle
I did not π¦ I tried to change code from <@141759663929556992> , but without luc...
In my case I was ended up using https://github.com/adafruit/Adafruit_CircuitPython_SPD1656 library for the 7color display. I think I tried a few other things before that, but this one Scott worked on before I had any hardware and it successfully displayed on my display without modification if I recall correctly.
slender iron
@vague thicket did you try with the regular il0371 driver?
I'd get two logic captures and compare them
vague thicket
<@340398038529081346> did you try with the regular il0371 driver?
There is some? I did not find one. I will capture data with my analyzer for Arduino and CPY and compare them, but I'm newbe with that as well.
slender iron
There is some? I did not find one. I will capture data with my analyzer for Ardu...
ah 0371
not quite the same
manic glacierBOT
blissful pollen
Lurking remotely today so did not add any notes
onyx hinge
https://blog.adafruit.com/2023/05/22/circuitpython-8-1-0-released/ https://blog.adafruit.com/2023/05/24/circuitpython-8-2-0-beta-0-released/
onyx hinge
sdomoszlai13 was on yesterday's report with a merged pull request, thank you!
and a new library that was on yesterday's report but not today's: jposada202020/CircuitPython_scales
idle owl
@midnight ember Thanks for handling the SSD1306 support issue!
lone axle
onyx hinge
for those who are curious the error that occurs when calling the show method with a Group is: ```Traceback (most recent call last):
File "code.py", line 41, in <module>
TypeError: function takes 1 positional arguments but 2 were given
Could we add specific error handing in `show` that prints a useful error?Here is the notes document for next Mondayβs CircuitPython Weekly meeting. It is at the normal time of 11am Pacific / 2pm Eastern here on Discord. Everyone is encouraged to attend! Please add your hug reports and status updates even if youβll be attending the meeting - itβs super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and weβll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/1v476RNCHysQUjWEPVojGsr7lcRD1saa2d9-Sdhbq3_Y/edit?usp=sharing
vague thicket
https github com adafruit Adafruit
midnight ember
Just making it back online. The SSD1306 thing I just wanted to bring up the issue. How it's handled is completely up to the librarians and learn guide people that is beyond anything else I could contribute to the conversation. Hearing the discussion amongst the people it's most relevant to was the goal for bringing up the issue.
Even something as simple as a header comment # this library is designed for framebuffio and not for displayio use or something as simple as that.
whatever you all decide how to handle the confusion is beyond me. had no idea it would affect so many other display libraries. if there's a larger naming convention confusion issue then it should be addressed across the libraries. that was a really interesting conversation.
tidal kiln
@tulip sleet is this still on going? specially - is there a known esp32-s3 specific i2c issue?
https://github.com/adafruit/circuitpython/issues/6311
midnight ember
part of the problem is when google searching circuit python ssd1306 examples it dumps you into an examples folder. you never see the base directory where the deprecated warning is. the deprecated page could have a huge warning with an animated gif myspace style and it would have no effect as the google search bypasses it and dumps you into the examples folder... which is what you want if you're searching for a github example. so the warnings are bypassed. header commenting in the example files is needed too.
midnight ember
<@126221640559362048> Thanks for handling the SSD1306 support issue!
Ultimately it was Neradoc that corrected the issue, I only provided an example for the user's question (once I used the right library).
idle owl
Ultimately it was Neradoc that corrected the issue, I only provided an example f...
I meant handling the long-term support issue by submitting an example. π
slender iron
@onyx hinge here is my brainstorm for what each physical synthio module would have: ```
Modular synthio
LFO:
OLED for displaying waveform
waveform: rotary switch with detents
rate: pot and input jack
offset: pot and input jack
phase_offset: pot and input jack
scale: pot and input jack
once: toggle
interpolate: toggle
output: output jack
Math:
operation: rotary switch with detents. 14 options
a: pot or input jack
b: pot or input jack
c: pot or input jack
output: output jack
Envelope:
OLED for displaying envelope.
attack_time: pot
decay_time: pot
release_time: pot
attack_level: pot
sustain_level: pot
output: jack
Noterizer:
OLED for waveform
waveform: rotary switch with detents
panning: input jack
amplitude: input jack
bend: input jack
ring_bend: input jack
winged willow
Has anyone got any resources for library construction? Specifically for i2c building?
Donβt really know what Iβm doing properly but Iβm trying to port a library for the Pimoroni IO Expander board.
Iβm getting some data out tho itβs not quite right yet, but Iβve a way to go and need to handle interrupts etc somehow too.
Hereβs what I have so far:
slender iron
I'd suggest starting with another circuitpython io expander library
onyx hinge
@slender iron cool! the number of analog I/Os required for some of those is large!
slender iron
None would be analog. The jacks would just connect one thing to another
Weβll pots would be analog I guess
still zephyr
Has anyone got any resources for library construction? Specifically for i2c buil...
Hello, I took a look at the code that you previously share. I would recommend use the Adafruit_register library, this will handle all the register address and masking and register for you, so you do not need to figure what mask to use. However, I am more a hands on person, so Ill try the code on the hardware. and then compared with other code available, I have done some I2C libraries from scratch for Circuipython. we could discuss if you need. my user handle is jposada202020 in github if you want to take a look. there are also some Circuitpython Expander libraries in the community bundle if you want to take a look https://github.com/adafruit/CircuitPython_Community_Bundle/blob/main/circuitpython_community_library_list.md. Good luck !
midnight ember
got a weird import order from pylint submitting an example ```py
Check failure on line 13 in examples/displayio_ssd1306_picowbell_tempsensor.py
GitHub Actions / test
third party import "from adafruit_bme280 import basic as adafruit_bme280" should be placed before "import adafruit_displayio_ssd1306 as ssd1306"
import time
import board
import displayio
import busio
import terminalio
from adafruit_display_text import label
from adafruit_bme280 import basic as adafruit_bme280
import adafruit_displayio_ssd1306 as ssd1306
``` if it wants 3rd party imported first then why doesn't it want it imported before time, board, etc..?the way pylint wants things formatted confuses the heck out of me sometimes, but i change it and it's happy so π€·ββοΈ
maybe OCD but I always do imports first and from second, it just looks cleaner. using Mu there's no import order helper.
midnight ember
Hello, I took a look at the code that you previously share. I would recommend us...
When using IO expander and multiplexer libraries, the register library is a common associated prerequisite so I would consider that completely normal if he goes with the register helper library.
jaunty juniper
the way pylint wants things formatted confuses the heck out of me sometimes, but...
it's alphabetical too, thus adafruit_bme before adafruit_displayio I think ?
midnight ember
is that all it is? lol i can remember that then. π
jaunty juniper
what pylint considers "3rd party import" I don't know though, are the CP builtin libraries defined in pylintrc ? I never thought about that
midnight ember
it only seemed to complain about the imported libraries in /lib yes
i've been using circuit python for like a year or two now and didn't know this...
i know it sounds really basic but can we get a learn guide on basic import ordering? might seem to be a nuance but not documented in any example to my knowledge.. π
crimson ferry
it's only relevant with pylint though, right? so, submitted code, maybe could go in the making libraries guide
midnight ember
That's true, Mu doesn't care about the ordering only pylint?
split ocean
<@426200547024961539> here is my brainstorm for what each physical synthio modul...
uh oh, he's gone modular!
edgy edge
I'm using a third-party machine learning library (Neuton). The vendor just stop shipping C source code and now supply .a and .h files instead. Is this the best approach to integrating with circuit python? It compiles and links successfully, but feels like a hack. Is there a better way to add an architecture dependent library? Is there a more convenient way to update OBJ to avoid the makefile edit.?```
Vendor supplied library files in lib/Neuton
neutonml folder with init.[ch] and neuton.[ch] in shared-[bindings, modules]
my board's mpconfigboard.mk:
AULITECH_NEUTONML = 1
py/circuitpy_defns.mk:
ifeq ($(AULITECH_NEUTONML),1)
SRC_PATTERNS += neutonml/%
INC += -I$(TOP)/lib/Neuton/neuton/include -I$(TOP)/lib/Neuton
LIB_3P += $(TOP)/lib/Neuton/neuton/lib/libneuton_arm_cm4.a
endif # note the library is architecture specific
ports/nrf/makefile:
$(BUILD)/firmware.elf: $(OBJ) $(GENERATED_LD_FILE)
$(STEPECHO) "LINK $@"
$(Q)echo $(OBJ) $(LIB_3P)> $(BUILD)/firmware.objs
adding LIB_3P to the list of object files
slight improvement ties the Library to board.```
my board's mpconfigboard.mk:
AULITECH_NEUTONML = 1
AULITECH_NEUTONML_LIB = lib/Neuton/neuton/lib/libneuton_arm_cm4.a
py/circuitpy_defns.mk:
ifeq ($(AULITECH_NEUTONML),1)
SRC_PATTERNS += neutonml/%
INC += -I$(TOP)/lib/Neuton/neuton/include -I$(TOP)/lib/Neuton
LIB_3P += $(TOP)/$(AULITECH_NEUTONML_LIB)
endif # note the library is now board specific
ports/nrf/makefile:
$(BUILD)/firmware.elf: $(OBJ) $(GENERATED_LD_FILE)
$(STEPECHO) "LINK $@"
$(Q)echo $(OBJ) $(LIB_3P)> $(BUILD)/firmware.objs
adding LIB_3P to the list of object files```
manic glacierBOT
onyx hinge
Weβll pots would be analog I guess
attiny seesaw has 9 ADC pins, or any old analog mux (e.g., 4051 family) would let you do up to 8
onyx hinge
https://blog.readthedocs.com/migrate-configuration-v2/
@idle owl @proven garnet this may need your attention across the libraries
Cookie cutter has a v2 file so maybe it's all ok
idle owl
<@330227457296957440> <@135415127879974912> this may need your attention across ...
I feel like we're good. But I'll take a look.
proven garnet
I feel like we're good. But I'll take a look.
Let me know if we need to address anything! And thanks @onyx hinge!
onyx hinge
Let me know if we need to address anything! And thanks <@426200547024961539>!
drivers/lsm303 has ".readthedocs.yml" while the standard seems to be ".readtehdocs.yaml". that one also lacks a "version: 2" line. The others all seem to have it. So, just one outlier. [in adafruit bundle]
onyx hinge
Let me know if we need to address anything! And thanks <@426200547024961539>!
in the community bundle, there are more outliers. https://gist.github.com/jepler/30114a9905feb4a9f0aee6ea137ca3e5 -- "biplane" seems to be incorrectly located, too
onyx hinge
Let me know if we need to address anything! And thanks <@426200547024961539>!
continuing to pile on, I noticed that there may also be deprecated nodejs actions in community bundle, there sure are in https://github.com/jepler/Jepler_CircuitPython_udecimal/ https://github.com/jepler/Jepler_CircuitPython_udecimal/actions/runs/5134723945
proven garnet
Oh wow these are good catches, I should have time to get to them in a little bit or worst case this evening!
onyx hinge
Oh wow these are good catches, I should have time to get to them in a little bit...
https://github.com/jepler/Jepler_CircuitPython_udecimal/pull/38 I tried to do my own package ...
I can't seem to invite you as a reviewer
since it's in my personal github
tulip sleet
@timid bolt now that 8.1.0 is done, I have finally gotten back to looking at asyncio. You'll see my comments in the PR's, but we could have a more general audio or video chat about it too if you have time.
proven garnet
https://github.com/jepler/Jepler_CircuitPython_udecimal/pull/38 I tried to do my...
I'll still leave a comment if you want if I see anything
onyx hinge
please do
slender iron
uh oh, he's gone modular!
it won't be eurorack compatible though
slender iron
attiny seesaw has 9 ADC pins, or any old analog mux (e.g., 4051 family) would le...
I'd want each module to be separate so I'd need 5 at most but could do that as two separate stemmas of 3 and 2
slender iron
slight improvement ties the Library to board.```
my board's mpconfigboard.mk:
...
Could you just OBJ += it from the board def?
manic glacierBOT
edgy edge
Could you just `OBJ += ` it from the board def?
I tried that an one point, but it looked like adding to OBJ triggered a make rule and caused a failure
idle owl
Oh wow these are good catches, I should have time to get to them in a little bit...
Thank you!
slender iron
I tried that an one point, but it looked like adding to OBJ triggered a make rul...
Ah, right. it tries to compile it. maybe a LIB def exists? I don't remember exactly
crimson ferry
sounds like c-stack
manic glacierBOT
proven garnet
drivers/lsm303 has ".readthedocs.yml" while the standard seems to be ".readtehdo...
This library is archived so we could fix it but have to unarchived it first. Otherwise if all of them have it, we should be good.
proven garnet
in the community bundle, there are more outliers. https://gist.github.com/jepler...
I could write a script to auto-PR these libraries, but I think for a while docs were not enforced for community libs (I may be mistaken though and thinking of PyPI related things). I suppose on some level of there aren't docs then if they ever get added then it can be fixed then.
proven garnet
Also took a look at your repository - looks like as long as there is a pyproject.toml file at all (which triggers the CI) the metadata is allowed to still be defined in setup.py.
slender iron
anyone know of a full implementation of uart using PIO?
PIOASM only has two separate examples
manic glacierBOT
manic glacierBOT
winged willow
Hello, I took a look at the code that you previously share. I would recommend us...
hey thanks a lot, that's awesome; I'll check that out right now...
I'm all kinds of useless with stuff like registers and byte shifting so a library that handles that sorta stuff a bit is great
onyx hinge
@proven garnet I'm wondering if there also needs to be a way to specify the --package_folder_prefix during the non-release build, build/action.yml. Because the artifacts I get from the jepler_udecimal build are just empty. I may be missing something of course
proven garnet
<@135415127879974912> I'm wondering if there also needs to be a way to specify t...
How are you checking it for build/action.yml?
I didn't even realize the non-release build had artifacts
onyx hinge
@proven garnet yeah you can grab the bundles from https://github.com/jepler/Jepler_CircuitPython_udecimal/actions/runs/5138862173 look inside one of the zips inside the zip (π) and see that it only has examples, no py or mpy files...
build/action.yml makes the bundle files here:76 run: circuitpython-build-bundles --filename_prefix ${{ steps.repo-name.outputs.repo-name }} --library_location . and then uploads them 78 uses: actions/upload-artifact@v3 but the resulting files are incomplete due to the same --package_folder_prefix problem .. but I don't see a way to speecify it, unlike at the actual release step
proven garnet
Oh interesting. Yeah, I think the mpy compilation is only patched in the release build, but it should be easy enough to add it to the non-release builds as you suggest
proven garnet
build/action.yml makes the bundle files here:``` 76 run: circuitpython-build...
Ah, I didn't realize there was an upload step! Is it worth adding it instead of removing the upload? I haven't thought about the use case of checking a build CI for release artifacts, but if there is one it's definitely worth updating.
I don't even think it would be a library patch - just patch the cookiecutter, find the few libraries that use it, and add the new info to them.
proven garnet
but since the fix is .. probably .. easy
I'm up for it! As far as I can grep only asyncio (and circuitpython_typing) technically need the fix
tulip sleet
I'm up for it! As far as I can `grep` only `asyncio` (and `circuitpython_typing...
neopixel too? or not because it is a single file library?
proven garnet
`neopixel` too? or not because it is a single file library?
Yup, safe because it's a single file!
I named the argument with "package" for that reason, but it's still somewhat confusing because it's still regarding "packaging"
lone axle
Anyone know where within the core to find the code that is responsible for changing the display back to displayio.CIRCUITPYTHON_TERMINAL when an exception has been raised? I.e. user has drawn something with displayio and it was successfully rendered, but then some time after that an exception is raised. By default the display goes back to showing the CircuitPython Terminal with the traceback from the exception being the last thing printed and visible.
lone axle
I may have worked that out. The display always gets reset after code.py is finished, whether or not it finished successfully or with an exception. During the reset it shows the terminal. If there was an exception it gets printed and can be seen on the display which is now showing the terminal. So there is nothing responsible specifically in the instance of an exception for changing the display it just happens because of the reset, which occurs whether or not there was an exception.
slender iron
I may have worked that out. The display always gets reset after code.py is finis...
That matches my understanding
slender iron
manic glacierBOT
timid bolt
<@1065334137243836517> now that 8.1.0 is done, I have finally gotten back to loo...
Hi @tulip sleet, good to hear from you. Does meeting 10 am PT tomorrow (Friday) or before/after the weekly meeting on Monday work for you?
tulip sleet
Hi <@329766224093249548>, good to hear from you. Does meeting 10 am PT tomorrow ...
thanks! - tomorrow or monday is fine. We have an internal meeting on Monday at 10am PT before the public meeting, so it would need to be before that
manic glacierBOT
timber mango
Can't get esp-idf to build within CircuitPython. Don't know if this is just me doing something dumb. I followed all the guides but there's still this weird error from Git.
Anyway, ping me if you have any ideas what's going wrong. Seems like Git can't retrieve the branch it wants?
brazen hatch
I suggest having a copy of espidf outside of your CP clone, and exporting that instead.
However if you want to fix the one inside your cp clone you should go inside the submodule and inspect with git status
timber mango
However if you want to fix the one inside your cp clone you should go inside the...
I'm getting the response to git status:
nothing to commit, working tree clean```I might try building seperately. What version should I use?:
The latest?
brazen hatch
The latest?
nope, 4.4.4 iirc (the latest 4.4.x release)
brazen hatch
I'm getting the response to git status:
```HEAD detached at 630c2724f
nothing to...
that is the intended output, idk
timber mango
Alright I'll try building from the 'releases/v4.4' branch.
brazen hatch
you could just git clone -b as they suggest in their release notes
brazen hatch
esp-idf ones
wraith crow
Can't get esp-idf to build within CircuitPython. Don't know if this is just me d...
After running make fetch-submodules the esp builds would sometimes fail for me. I don't know if you're running into the same issue but for me running a make remove-submodules and then re-running the make fetch-submodules would resolve the error. I recently moved my build machine from a WiFi SBC to a wired desktop and don't think I've seen the issue since.
proven garnet
@vocal badger Brent, I think this is you, you asked me to reach out to you but wanted to make sure this was the correct handle.
timid bolt
hey @tulip sleet, do you want to chat now?
manic glacierBOT
tulip sleet
hey <@329766224093249548>, do you want to chat now?
whoops, sorry, I got interrupted, sure, that's fine
timid bolt
all right, which room?
tulip sleet
amelia earhart is fine
tulip sleet
@timid bolt https://discuss.python.org/c/async-sig/20
manic glacierBOT
tulip sleet
manic glacierBOT
timber mango
After running make fetch-submodules the esp builds would sometimes fail for me. ...
Okay, I'll rerun the command to make the submodules.
But it's a fresh build that I haven't modified at all.
wraith crow
Yep that was when it happened to me as well
timber mango
Ah, well I don't understand why. But if it works, I don't have to know! π
I was hoping to make some small modifications to my firmware. Namely including the audioio module.
wraith crow
That was me as well π
tulip sleet
audioio depends on the existence of a DAC, so it may not work on the board you have in mind.
timber mango
It works on the ESP32 right?
Just tell me I'm not doing all this only to run into some unsolvable technical problem? Right? π
(but I have tested that sending a pwm signal to the pin will produce the correct frequency on the speaker)
So do I need audiopwmio instead?
wraith crow
Didn't Adafruit just add a little add on board with i2s out?
tulip sleet
audioio is only implemented for atmel-samd
audioio should really be called audioanalogio. Yes, you could use audiopwmio or audiobusio (I2C) if available
timber mango
Okay so what I want is audiopwmio then. I already have audiobusio, it's just not very helpful as I don't have a digital audio device.
So after fetching the submodules I need to run 'esp-idf/install.sh'... I think?
tulip sleet
right
then, before you do a build, run esp-idf/export.sh
that sets up the right PATH, etc.
timber mango
Still the same problem.
Checking "python" ...
Python 3.10.6
"python" has been detected
Installing ESP-IDF tools
Current system platform: linux-amd64
Selected targets are: esp32s2, esp32h2, esp32c3, esp32s3, esp32
Installing tools: xtensa-esp-elf-gdb, riscv32-esp-elf-gdb, xtensa-esp32-elf, xtensa-esp32s2-elf, xtensa-esp32s3-elf, riscv32-esp-elf, esp32ulp-elf, openocd-esp32
Skipping xtensa-esp-elf-gdb@11.2_20220823 (already installed)
Skipping riscv32-esp-elf-gdb@11.2_20220823 (already installed)
Skipping xtensa-esp32-elf@esp-2021r2-patch5-8.4.0 (already installed)
Skipping xtensa-esp32s2-elf@esp-2021r2-patch5-8.4.0 (already installed)
Skipping xtensa-esp32s3-elf@esp-2021r2-patch5-8.4.0 (already installed)
Skipping riscv32-esp-elf@esp-2021r2-patch5-8.4.0 (already installed)
Skipping esp32ulp-elf@2.35_20220830 (already installed)
Skipping openocd-esp32@v0.11.0-esp32-20221026 (already installed)
Installing Python environment and packages
fatal: No tags can describe '630c2724fc8c69eeaaa1bb025de52b99c5cb11aa'.
Try --always, or create some tags.
WARNING: Git describe was unsuccessful: b''
ERROR: This script was called from a virtual environment, can not create a virtual environment again
Hmm, well maybe I can try building my own version of esp-idf from the 4.4 branch of their GitHub?
tulip sleet
if git status shows modified submodules, you need to go into each submodule and do git restore .
tulip sleet
Hmm, well maybe I can try building my own version of esp-idf from the 4.4 branch...
that should not be necessary at all. another simple thing is just to clone again and do make fetch-submodules on the fresh clone
timber mango
git status says it's clean.
Your branch is up to date with 'origin/main'.
nothing to commit, working tree clean
can also try make remove-submodules and then make fetch-submodules. But the re-clone is easiest unless you will lose some work
when submodules move around or change, git can get pretty confused
timber mango
I was already using a fresh clone, and I just ran make remove-submodules and make fetch-submodules.
I guess I could burn it all and try again?
tulip sleet
what is the host OS, and what is the version of git you are using?
tulip sleet
another idea: isudo apt install git-lfs. This was missing from the instructions. I just added it -- sorry. THen do the remove and fetch again, or a fresh clone. I have added a new page just today https://learn.adafruit.com/building-circuitpython/recent-changes
tulip sleet
yes, you are using an older version of git than I am, but same OS. git 2.36.0 is the minimum required to do "partial clones", but you don't really need those
I am working on make fetch-submodules variations that will be only for a specfic port, to avoid having to fetch tons of useless stuff
timber mango
I can try updating git? I'll grab a snap package or something similar. I expect the distro version is pretty old.
tulip sleet
I use the git ppa: https://launchpad.net/~git-core/+archive/ubuntu/ppa
but a staff member had the same problem you are having earlier today (on macOS, though), and installing git-lfs fixed it
timber mango
Okay. I've added the 'ppa:git-core/ppa' repository. I'll remove and fetch the submodules again.
git version 2.40.1
That would be pretty useful right about now, haha.
Alright that the submodules fetched. Seems to be working a little differently now. Git was throwing errors about things being not a tree before. So the lack of errors now gives my some reason for optimism!
Ah... crap. ```Detecting the Python interpreter
Checking "python" ...
Python 3.10.6
"python" has been detected
Installing ESP-IDF tools
Current system platform: linux-amd64
Selected targets are: esp32, esp32s3, esp32c3, esp32s2, esp32h2
Installing tools: xtensa-esp-elf-gdb, riscv32-esp-elf-gdb, xtensa-esp32-elf, xtensa-esp32s2-elf, xtensa-esp32s3-elf, riscv32-esp-elf, esp32ulp-elf, openocd-esp32
Skipping xtensa-esp-elf-gdb@11.2_20220823 (already installed)
Skipping riscv32-esp-elf-gdb@11.2_20220823 (already installed)
Skipping xtensa-esp32-elf@esp-2021r2-patch5-8.4.0 (already installed)
Skipping xtensa-esp32s2-elf@esp-2021r2-patch5-8.4.0 (already installed)
Skipping xtensa-esp32s3-elf@esp-2021r2-patch5-8.4.0 (already installed)
Skipping riscv32-esp-elf@esp-2021r2-patch5-8.4.0 (already installed)
Skipping esp32ulp-elf@2.35_20220830 (already installed)
Skipping openocd-esp32@v0.11.0-esp32-20221026 (already installed)
Installing Python environment and packages
ERROR: This script was called from a virtual environment, can not create a virtual environment again
I think that's a known bug with the ESP-IDF build.
tulip sleet
Alright that the submodules fetched. Seems to be working a little differently no...
so you are running Python in its own virtual environment? I use the system one and don't see this.
timber mango
No I am not running Python in a virtual environment.
But I found this post in the ESP-IDF Github issues:
https://github.com/espressif/esp-idf/issues/8470#issuecomment-1477324664
tulip sleet
are you using the submodule version of ESP-IDF in CircuitPython, or is this some separate thing?
timber mango
It seems there was an unrelated virtualenv in my PATH from an old PlatformIO install. Not active, just in the path. Removing it fixes the problem.
Excellent, now to build.
timber mango
Alright the build is successful. Time to mod.
So if I'm reading this documentation correctly, all I need to do is to add CIRCUITPY_AUDIOPWMIO = 1 to the specific board's 'mpconfigboard.mk' file?
Is that really it?
Ah, maybe not.
that might be because you need a make clean ?
timber mango
I ran make clean from the top level directory. Is that correct? Or there some different version I need to run for the espressif directory?
tulip sleet
Alright the build is successful. Time to mod.
So if I'm reading this documentati...
which board is this?
timber mango
M5Stack Core Basic (m5stack_core_basic)
I ran make BOARD=m5stack_core_basic clean from the espressif directory, but still the same issue.
That's an ESP32 based board.
tulip sleet
I ran `make clean` from the top level directory. Is that correct? Or there some ...
run make clean in ports/espressif
timber mango
I did run make BOARD=m5stack_core_basic clean from that directory.
(it doesn't seem to work without specifying the board)
tulip sleet
if it were, it probably already be turned on. We don't generally turn things off unless the board is extremely small and it wouldn't fit. ... So you need an I2s amp
no
you can use PWMOut to make tones, but nothing fancier than that
timber mango
Ah, well I guess it's a dead end then. I'll just have to convert my compositions into an array of PWM frequencies or something like that. Maybe I'll write a script for that. Anyway, problem solved.
tulip sleet
timber mango
I know, but I'd probably need to design my own module to mod it to the existing board.
timber mango
To explain, this is a unit contained within a case. I can't just stick a module on.
tulip sleet
does it already have some kind of audio? Was it running Arduino or something?
timber mango
Yeah, Arduino and MicroPython I think. Though I think it was just doing tones there as well.
I might just have to update to the newer version of the board which does have an I2S amp in the schematic.
Anyway, gotta go now, but thanks for all your help! It's been really useful in getting my build working.
tulip sleet
sure! sorry it was disappointing. The DAC audio was not really working well (or was not usable) in ESP-IDF when we first looked at it.
lapis shuttle
hello friends, does anyone know the difference between MP_OBJ_NEW_QSTR and MP_ROM_QSTR in circuit python port/$arch/$board/pin.c? was trying to port circuit python to a custom board with different pin layouts but confused which one should I use for my pins
manic glacierBOT
manic glacierBOT
manic glacierBOT
upper ocean
Hi folks, I'm hoping to spend some time digging into https://github.com/adafruit/circuitpython/issues/7966 to see if I can find a way to enable native code emitters. Since I'm using an esp32s3 (UM FeatherS3, specifically), I was hoping to using the builtin JTAG for live debugging to see what's going on when CP crashes.
However, all my attempts at using openocd and gdb with this board have failed. For the devs here, I'd love to hear what kind of debug processes you use, especially for developing with the esp32s3. Anyone have a reliable way to live debug w/ gdb? What about getting backtraces from crashes?
Thanks!
manic glacierBOT
brazen hatch
Hi folks, I'm hoping to spend some time digging into https://github.com/adafruit...
c3 is the only esp you can debug without burning fuses currently
CP switches out the jtag for the Circuitpy drive on S3
I want to get that changed on DEBUG eventually
manic glacierBOT
tulip sleet
i have had very little success trying to do use gdb with a J-Link on the ESP boards. It may sort of work for a while, but gets stuck, or is very limited. Instead we have compiled DEBUG=1 and used the ESP_LOG*() macros to see debugging printouts on a designated set of UART pins.
left iris
Hi newbie here. This was my enhancement request. Now that I have the board I might as well contribute to add one other board. Is there a guide that adafruit has on how to add a new device for circuitpython support? I know it has micropython support cause the main page provides a build. And the device exists in their repo.
I was building my project on the Pico W - ran out of memory sigh and so would be grateful if someone could help π
tulip sleet
@left iris ^^
winged willow
Hey all. Iβm still working on my library to get the Pimoroni IO Expanded Board (Nuvoton ms51 chip) working in circuitpython
Run into a glitch tho - Ive managed to get regular IO pins working - I can see input on pull-up pins etc when theyβre switched on etc etc BUT 2 pins donβt work at all. Others do so I canβt see what Iβm doing wrong. Iβve checked the registers for those pins against the micropython and arduibo libraries and theyβre correct. Is my code wrong or is my board toast..
Hereβs the library:
https://github.com/mathcampbell/IOExpanderCircuitPython
upper ocean
I want to get that changed on DEBUG eventually
Thanks for the info, even if it's non-ideal. I noticed with Arduino there is the option to select USB Modes like "Hardware CDC and JTAG" as well as "MSC". I assume that CP is using the "MSC" option, or something like it, to expose the CP drive, whereas the "Hardware CDC and JTAG" option, aka Serial JTAG, is what we'd want for debugging. What I can't figure out is what settings Arduino is changing when a user selects one or more of these USB Modes. Do you know if it is just a config change like in sdkconfig? It'd be great if the only thing that needed to be done to temporarily disable the CP drive and enable JTAG was to change a couple values in sdkconfig.
brazen hatch
Thanks for the info, even if it's non-ideal. I noticed with Arduino there is the...
Nope it's actually code changes.
manic glacierBOT
slender iron
Hey all. Iβm still working on my library to get the Pimoroni IO Expanded Board (...
Do they work in the MP library?
slender iron
Thanks for the info, even if it's non-ideal. I noticed with Arduino there is the...
I've never had a lot of luck using a debugger when timing sensitive protocols like BLE and Wifi are running. I too do printf debugging on ESP. During faults, the panic handler will print out a backtrace for you. You can also add backtrace prints for yourself if you want to too.
winged willow
Do they work in the MP library?
thanks; realised soon after I should check that so I did, also checked using the arduino library someone did. WOrks fine in them so it's my code that's to blame...
slender iron
A logic analyzer could help you determine what the driver difference is
winged willow
seems to be i can read ll the pins, except those where the port needs to be acssed with the 0th bit
gilded cradle
@slender iron I need to take care of something that happens the same time as the meeting. Could you read my notes?
slender iron
<@252717193496756235> I need to take care of something that happens the same tim...
yup! no problem
gilded cradle
yup! no problem
Thank you
onyx hinge
<@&356864093652516868> hey y'all! We'll be holding the weekly meeting in about 30 minutes -- now's a great time to finish up your notes in the notes doc: https://docs.google.com/document/d/1v476RNCHysQUjWEPVojGsr7lcRD1saa2d9-Sdhbq3_Y/edit?usp=sharing
slender iron
thanks, meant to do that
digital ibex
How is the "In the weeds" section structured? I know it's a place to put other topics of interest, but do I put things in there stated as questions?
slender iron
we read each item and then respond/discuss
I'll either read the topic or handoff to the poster if they are in the meeting
digital ibex
Ok, thanks for clarifying!
slender iron
np π
idle owl
Ok, thanks for clarifying!
I formatted it a little differently for you. It's usually easier to keep track of things with that formatting, especially if there are many topics. π
digital ibex
Thanks, it does looks better now!
onyx hinge
@slender iron kattni asked me to read the libraries section today (also noted in the doc)
@slender iron about ready to kick things off?
oh no we're going up against the apple keynote? π°
digital ibex
Looks like I'll be missing the rest of this meeting, sorry. Can someone read off my in the weeds question?
Thanks!
slender iron
I can read it
digital ibex
I'm going to listen in, but I'm text only
lone axle
It's a displayio based "emulator" that draws on a display to mimic 7segments, 14segments, or 8x8 matrices.
onyx hinge
@lone axle ooh cool!
[hum should we put the required mpy version in the banner or boot_out.txt?]
slender iron
weeds topic?
ember iris
Oh those filters take me back
digital ibex
I missed most of what Dan said but I'll look at the recording. Docker is basically a reproducible VM.
digital ibex
That's a fair point about USB, I wonder if there's a better container solution or something else to make it easier. Thanks for the input!
blissful pollen
If you want to run anything past me about it let me know. If I'm really stuck I can probably ask someone I know for a hint
turbid radish
ha ha
digital ibex
π€£
But it would need to exist on a bunch of CDs for the authentic Microsoft experience!
ember iris
I do think if skipping a 9.0 full release makes the process of maintaining code easier, then it's worth skipping after an alpha
onyx hinge
rationale: we don't know how much existing software checks for "9" as the first digit of the number and assumes it's working with windows 95/98 and not circuitpython
Next meeting is at the normal time on Monday June 12!
turbid radish
Windows Forever
digital ibex
π§π§
onyx hinge
Python 3.11 for workgroups
digital ibex
Bye all!
onyx hinge
π
lone axle
Thanks for hosting Scott. Have a great day everyone π
ember iris
Thank you all! Have a great week!
onyx hinge
If hypothetically the docker build looked like sh tools/docker-build espressif adafruit_matrixportal_s3 i'd be super into it because i wouldn't have to worry about anything but getting docker set up initially (and having enough resources)
brazen hatch
If hypothetically the docker build looked like `sh tools/docker-build espressif ...
It wouldn't be all that hard to configure really. Just slapping pwd as a bind mount and and an env for the board should be all that it takes.
If nobody does it, I will do it eventually, now stuck doing telnet.
I was fumbling around with the chinese w801 chip yesterday, and lost a decent bit of time.
I got uart, wifi and gpio going. The docs are basically non-existent, but I am a tryhard, so did it without any.
I got heap (malloc's) and stack (static allocations in code) going, along with freertos task scheduling.
I wonder how much effort would it take to port circuitpy to it.
The functions it has for all the hardware are very simple, and just work.
Also that chip goes like 60Β°C and draws .12A just doing some leds.
slender iron
Here is the notes document for next Mondayβs CircuitPython Weekly meeting. It is at the normal time of 11am Pacific / 2pm Eastern here on Discord. Everyone is encouraged to attend! Please add your hug reports and status updates even if youβll be attending the meeting - itβs super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and weβll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/1GiH22DPaCVQ8D4cjhq2yJR73_bm3sTRZ35IE_L2bFBc/edit?usp=sharing
slender iron
Here is the notes document for next Mondayβs CircuitPython Weekly meeting. It is...
onyx hinge
@slender iron this layout of holes gets all the multiples of 0.2" and also allows placing feathers with holes on an 0.7 x 1.8" grid at a variety of locations and orientations. you could do the overlapping holes as shown or as slots.
assuming I mathed correctly anyway
slender iron
oh, fancy
I was thinking about surveying our existing designs
that's a very cool design
winged willow
urgh I'm beginning to hate this library lol.
2 weeks i've been smashing away at it. I can read some pins but not others..the pin addresses are correct, Iv'e checked the hardware in the arduino and micropython version...
i even had chatgpt sit and compute the binary for me for each i2c write etc...
Anyone care to take aa look and see where i've gone wrong?
https://github.com/mathcampbell/IOExpanderCircuitPython
slender iron
you'll need to provide more specifics. that's 500 lines of code
and chatgpt makes stuff up
winged willow
hahm, fair, I didn't wanna launch into a whole explanataion if nobody was wanting to
yeah..it broke off from exaplining the hex-binary conversion to tell me about president of france
so at line 219, the _set_mode method
the pin map showing pins: port/bit, taken and cross referenced from the arduino and micropython versions
that uses that to decide which registers to address in the setting of the io mode for each pin (input/pullup/adc/input//output etc etc)
thing is, pin's 2, 5 and 10 do NOT work. At all.
The rest, do.
The only commomon linking them is that pins 2, 5 and 10, you have to use their port, 0 bit.
so I'm thinking the code that assigns that particular pins is doing something bitwise and adressing the wrong bit or something, expect that the code looks right...
i got one erroe message suggesting those pins may even be getting set to output instead of pullup
winged willow
no; i'm just trying to get the general IO mode going for starters...
slender iron
line 266
winged willow
haven't even really touched that code yet...
orchid basinBOT
slender iron
302 doesn't look right since bit doesn't shift anything
winged willow
bypassing all of that to just get IO pins working since on this board all pins can be IO (and are by default), and some are also PWM or ADC
a lot of the code is my just copying the (working fully) arduino library;
//5th bit of mode encodes default output pin state
i2cWrite8(ioPin._reg_p, (initialState << 3) | ioPin._pin);
slender iron
do you have a datasheet to look at?
winged willow
pin in that librbary fo some reason was waht he confusingly labelled not only the actual pin but also the bit to address on teh in pin
yes and no
but i'm not sure it's entirely accureate cos Pimoroni reference them having their own custom firmware, and it talks about some registers being shifted...
slender iron
right, so it's their version of seesaw
I'd use a logic analyzer to compare the implementations' i2c traffic
winged willow
i'm also getting from the print status messages that my schmitt trigger isn't actually assigned, but i've double triple checked and i'm using the exact same registers and bit operations etc as the other libraries, but it's not writing them for some reason...
slender iron
the other thing I'd do is slim the code down to just one thing you are trying to do
winged willow
i haven't got one (and wouldn't know hopw to use one if I did tbh, I'm not a programmer really..)
slender iron
pimoroni are the folks to ask for support since it is their hardware and firmware
winged willow
frankly i'm amazed I've even gotten this working enough to actrually get any data out of it...the other pins work..
i only got the board coes I'm building a joystick and it fits neatly in te space I had my mcp23017 but has some adc on it..
slender iron
I was expecting to find an if bits: but didn't see any
winged willow
where?
slender iron
in your code since the value 0 evaluates to false
winged willow
no i meant in what section
slender iron
any of them
winged willow
is that a convention?
I really dont' do much codding otehr than editing other people's, and python of any kind is kinda new to me, I learned a bit of objC years ago then some c++ etc
bitwise stuff is just alien to me...
slender iron
that's something that could explain why non-zero bits work
winged willow
this code is literally just me trying to copy the arduino and c++ code
with "what does this mean" questions at chatgpt lol
it is quite amazing i've somehow accidentally managed to not only read some data out of it but it even works...I can reliably read an io input pin 1-14 except those whose port needs 0pth bit addressed
the annoying thing tho is i'm moderately sure I basically understand how to get data out, and I'm positive I've copied the arduino code perfectly:
you get the bit using the Px registers array, corespoinding to teh port you're acessing; and the value is found in the corect bit of the returned string
slender iron
pimoroni may be better to ask since they know the firmware
winged willow
yeah I'm thinking so...I even checked the logic thru etc, and it seems to be "right" but trying to write the register for the schmitt trigger doesn't seem to "take"
which suggests to me something else is wrong (all likely tied to the 0th bit being read from the port)
i just realised I'm not setting bits using the special case registers it mentions///
probaly wouldn't explain why the pins don't work propelry buit might explain why they won't pullup
onyx hinge
@slender iron I made a slotted version of that layout and printed it .. a feather fits. the stemma board I grabbed to test with is https://learn.adafruit.com/adafruit-stemma-non-latching-mini-relay/downloads which is NOT on the 0.2" grid but the slots allow it to be screwed in at 2 positions which is also better than what a simple 0.2" grid of holes allows.
winged willow
seem to have fixed the issues with my library. The write function I was using was not constructing teh bytearray properly so that it weas only sending 1's but not 0's...
yes, I know i suck at coding
orchid basinBOT
slender iron
seem to have fixed the issues with my library. The write function I was using wa...
Nice! That's how you learn. Last week I spent a non-trivial time debugging what turned out to be swapped wires.
winged willow
yep it sucks doesn't it
slender iron
all part of the game π
winged willow
Given itβs only my second time doing any kind of circuitpython (any kind of Python in fact) coding, and the first time was my joysticks βif this pin reads true, so the joystick button of the same nameβ simple stuff, Iβm quite smug
slender iron
it's a good feeling to get things working
winged willow
Admittedly the adc function doesnβt work yet but Iβll get there. It was this or buy a brand new board at 3x the price and have got change my 3d printed joystick design to fit it in.
Plus it also means other people can now use this board and not have the pain I have had since I put it on GitHub
manic glacierBOT
same problem with 8.2.0-beta.0. After hitting the return on the 'import wifi' in the REPL, REPL locks up, and garbage characters also show up on the ODroid Go screen (see attached pix of screen before and after hitting the enter key). Similar to screen in issue #8069 and #6936.
Before Enter key:
After Enter key:
 The instructions in the guide have been updated: https://learn.adafruit.com/building-circuitpython/build-circuitpython#fetch-submodules-3141425. Also note there now is a "Recent Changes" page in the Building CircuitPython guide: https://learn.adafruit.com/building-circuitpython/recent-changes
slender iron
@vale spire have you looked into stack size optimization at all? our copy of mp_execute_bytecode takes just over 200 bytes a call
viscid pine
can i get the circuitpythonista role please
lone axle
can i get the circuitpythonista role please
got it added for you.
viscid pine
thankyou! π
still zephyr
<@&356864093652516868> **ALL FOLKS who do CircuitPython builds:** as of today, <...
π had the following error for the raspberrypi port
fatal: Unable to checkout '2001c62d1a0361c44acc7076d8ce91e1d1c66141' in submodule path '../../tools/python-semver' fatal: Unable to checkout '1421ca7fe72a2b173b08be05f9042b79409e783a' in submodule path '../../tools/uf2' Entering '../../data/nvm.toml' error: cannot run data: No such file or directory fatal: run_command returned non-zero status for ../../data/nvm.toml
full log here https://gist.github.com/jposada202020/f0f70d561715e20f8abf8ffdd134732e. This is a fresh cloned repo. I have tried before with an old one doing make remove-all-submodules same thing, I am in ubuntu, maybe doing something wrong...idk
tulip sleet
π had the following error for the raspberrypi port
`
fatal: Unable to checkout ...
what is git --version
still zephyr
2.34.1
tulip sleet
i'll bet this is a git version issue. I will put that version in my path and debug. 2.40 (2.36 or later, really) should work. You can use the git ppa if you want to get a later version.
thanks for testing that
still zephyr
will do, will need to investigate how to add the ppa because keep forgeting :).
tulip sleet
will do, will need to investigate how to add the ppa because keep forgeting :).
https://launchpad.net/~git-core/+archive/ubuntu/ppa ; the instrutions are right near the top of the page
still zephyr
Thanks doing that atm, I found your previous instructions here π
tulip sleet
are you on ubuntu 22.04?
tulip sleet
yes ubuntu 22.04
great, I will need to fix this for people who can't upgrade,
still zephyr
Thanks dahn!
manic glacierBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
lone sandalBOT
tulip sleet
@slender iron should I make an 8.1.1 release to fix the CircuitPlayground issue with stack overflow? The changes are pretty significant, though.