how are you connected to serial?
#circuitpython-dev
solar whale
I've tried screen, rshell, thonny
solar whale
in konsole I had to add the title bar to the tab title format
hmm -- I'll take a look for something like that.
slender iron
Adafruit CircuitPython 8.0.0-alpha.1-91-g0c46e990f-dirty on 2022-08-03; ESP32-S3-USB-OTG-N8 with ESP32S3
>>> import wifi
>>> wifi.radio.ipv4_address
192.168.1.94
next thonny beta will show it
solar whale
>>> import wifi
>>> wifi.radio.ipv4_address
10.0.0.158
>>>
``` that s a good sign....ooops wrong board
jaunty juniper
there's something, I saw the title bar saying "no IP" or something, but now that it's connected the C3 or the FunHouse give me no title bar in tio
(when they used to)
slender iron
hrm, could be a bug. I've got more changes pending for it
solar whale
>>>
>>>
>>> import wifi
>>> wifi.radio.ipv4_address
>>>
``` nothing for C3>>> os.listdir()
['.fseventsd', '.metadata_never_index', '.Trashes', 'code.py', 'lib', 'boot_out.txt', '.env']
>>> ``` same as on my other boardsand the .env is the same too?
solar whale
yes
jaunty juniper
when is the title bar sent ?
if I disconnect from the board, then reconnect from a new shell, what does it take to get the title bar ? ctrl-C ? ctrl-D ? a new prompt (enter) ?
devout jolt
I’ve noticed sluggishness in the REPL on reset as (I think) the supervisor is trying to connect to WIFi with bad credentials or weak signal. After a minute or so when it fails, sluggishness is gone
solar whale
I just verified that the .env is correct (same as working boards) could be a weak signal issue -- maybe the qtpyc3 is less tolerant -- I'll try moving it cloaser to the router.
slender iron
@jaunty juniper want a build from my working branch?
jaunty juniper
meh, I'll see when it's merged, I can't spend too much time on C3 testing rn
solar whale
moving closer to the router did not help 😦 seemed too easy
spiral elk
Wonder if it has the same need to set txpower as the Lolin C3 mini
I needed this in boot.py on the C3 mini in addition to the .env file stuff to get web workflow up and running:
import wifi
wifi.radio.enabled=True
wifi.radio.tx_power=8
@slender iron aha -- moving closer to the router did make a big difference -- it works fine when closer.
My signal in my "office" is rather weak but most boards are OK -- some are problematic -- looks like the qtpyc3 qualifies.
@spiral elk I will try that.
@spiral elk at first attempt, that did not appear to help, but I'll keep trying -- location does matter!
looks like I need to try my wifi extender again....
slender iron
I wonder how I could detect that. I found some good Unicode bars for showing signal strength
solar whale
Adding a "mesh pod" to my xfinity router seems to make the C3 much happier in my office. -- Now I just need to see how the overall network performs -- I have had issued with the xfinity pod in the past, but it's worth another try. Anyway CP is exonerated...
hidden rain
Not trying to grandstand, but thought I might shared my first 'toy' extension of circuitpython (ala everyone's gracious help). One can find the CPy code at https://github.com/latkinso42/fibonacci and the extension branch at https://github.com/latkinso42/circuitpython branch fib. The README has the details of usage. This was done in preparation of an ADC through DMA project to learn the ins and outs of extending circuitpython. Comments are Welcome. Thanks!
onyx hinge
Not trying to grandstand, but thought I might shared my first 'toy' extension of...
@hidden rain you should submit this for the newsletter, looks cool.
To contribute your own news or project, edit next week's draft on GitHub and submit a pull request with the changes. You may also tag a tweet with #CircuitPython on Twitter, or email cpnews@adafruit.com.
@lone axle hey if you get a chance can you create the notes doc for next week?
slender iron
icymi: ppk2s are in stock at digikey: https://www.digikey.com/en/products/detail/nordic-semiconductor-asa/NRF-PPK2/13557476
hidden rain
<@314934150647906305> you should submit this for the newsletter, looks cool.
> T...
Thank You. I'll try to get (or git) to that tonight.... 😉
lone axle
<@382939733107408897> hey if you get a chance can you create the notes doc for n...
whoooops, yep will do.
Here is the notes document for next Monday’s CircuitPython Weekly meeting. It is at the normal time of 11am Pacific / 2pm Eastern here on Discord. Everyone is encouraged to attend! Please add your hug reports and status updates even if you’ll be attending the meeting - it’s super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and we’ll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/1Zm1TPFm0mBhFvzQWCashMrcrBoR_d2SQMGtBrvQ7oSE/edit?usp=sharing
exotic pumice
Thank You. I'll try to get (or git) to that tonight.... 😉
git 'er done 😛
onyx hinge
whoooops, yep will do.
thank you!
solar whale
@slender iron just FYI -- all of the intermittent behaviors I was having with web workflow seem to have cleared up by improving the wifi signal... It is not happy with weak signals and there are clear differnences between boards as to what they will tolerate.
hidden rain
@onyx hinge You was suggested I contribute my little circuitpython extension adventure to the newsletter. I have a mini article (~ 50 lines). Is that too much? Most of what I see in the news letter is very free flowing (not a traditional article format). This may not be the channel for these questions...
onyx hinge
@hidden rain usually the newsletter is a few sentences, a photo, and a link to more info. If you don't blog or use twitter, you can do something like create a github gist with the "long version" and use that as the link.
hidden rain
ok. I will reduce to a few sentences, then just edit the 08-9-22 draft, then request the pull (submission) . ok.. all new and exciting.
random junco
ok. I will reduce to a few sentences, then just edit the 08-9-22 draft, then req...
I send in a PR every week for the podcast, Anne is great to work with on the newsletter
slender iron
anne is out today and tomorrow for the most part
random junco
oh yeah, she's getting married!
hidden rain
OK. A title and four lines referencing the full story on my GitHub.
random junco
OK. A title and four lines referencing the full story on my GitHub.
Nice! It'll be a few days probably until Anne has a chance to review it
hidden rain
ok.. Hopefully I did it correctly!
slender iron
welp, I got samd21 building with my libgcc that is smaller
can I "dibs" those bytes?
blissful pollen
Finders keepers seems to apply here
tulip sleet
can I "dibs" those bytes?
That's what I did 😆
slender iron
fills up the CI
onyx hinge
welp, I got samd21 building with my libgcc that is smaller
Only when the GitHub ci does the same
slender iron
Only when the GitHub ci does the same
I made the CI use it
onyx hinge
I made the CI use it
Yay!!
vale spire
heyya, how does CircuitPython currently handle "user wants to extend builtins from python" (i.e. the thing we currently do with the u-prefix)
(and related, how would you like it to work if we were to fix it upstream)
slender iron
heyya, how does CircuitPython currently handle "user wants to extend builtins fr...
No, we'd inherit it from you if you did
vale spire
ok. how do you currently support it?
or do you provide more stuff by default to make this less important? (os.path is a simple example off the top of my head)
slender iron
or do you provide more stuff by default to make this less important? (os.path is...
we don't provide it and haven't heard too many requests from folks
manic glacierBOT
vale spire
we don't provide it and haven't heard too many requests from folks
ok thanks. just raised https://github.com/micropython/micropython/issues/9018 if anyone is interested in this topic
manic glacierBOT
lone axle
With regards to the library size discussion from weeds of this weeks meetings... I'm tinkering with some utility to clone a repo / branch, build mpy files from it, and then trying to output the most helpful measurements from that (and eventually intend to attempt loading it on a device, but not there yet).
Someone mentioned running strings some_lib.mpy to get output of all of the strings within the library that will consume RAM. Would it then be helpful to do something like strings some_lib.mpy > strings_output.txt to save that to a file and then output the size of that file? Or would it be better to somehow measure it without explicitly writing to a file? I'm not sure if that would have any impact on the measurement.
jaunty juniper
With regards to the library size discussion from weeds of this weeks meetings......
be careful that MPY files include the path you provided to mpy-cross, unless you use -s to override that
(that will affect the file size)
lone axle
I'll have to look into how build-tools handles it. I'm using this to build instead of mpy-cross directly:
circuitpython-build-bundles --filename_prefix {get_project_dir_name(REPO_URL)} --library_location .
ok so there's no issue then, it should give it the correct path
lone axle
Nice, thank you for the heads up. I'll double check my built versions against the one published in the bundle as well to ensure they are the same.
I think it shouldn't be too hard to have an actions task that can measure these things and output it as a comment on PRs or somewhere in the actions output info. That can give us at least a baseline quantification.
I have bigger and more complex plans to also attempt setting up a system that could load it on a device and take measurements from gc.mem_alloc() before and after importing and output those as well.
hidden rain
To extend circuitpython to capture ADC using the DMA to a 'one-shot' buffer, one could extend the current AnalogIn with optional arguments or create a new interface AnalogDMAIn. The first is more universal and theoretically optional arguments would protect existing usage, yet the second although less universal forces a deliberate choice to use a new interface but at the cost of entirely new files. Please comment on the preferred choice for users since this could potentially be used across all platforms with a DMA & ADC. Thanks.
stuck elbow
one problem is that it would need to be optional and disabled on the platforms that don't have enough flash for the extra code
that is easier done with a separate module
it's also easier to explain to the users that this functionality is missing then
hidden rain
The intent, at first is to create this for the RP2040 (Pico), not ambitious enough for multi-platform at this point. However, I see that it could be extended for other platforms. I think the same module, namely analogio could contain the new code, but is see the issue as altered current interface verses a new file(s) for a a new API per platform.
So if the platform does not have DMA or enough flash, then the file(s) would not be extended for that platform.
This is largely an architectual issue, right?
stuck elbow
I think that all boards supported by circuitpython have dma
there is a similar, but separate, issue of making all the input/output modules compatible with asyncio somehow
slender iron
generally we disable and enable functionality at the module level so I'd suggest a new module
stuck elbow
I wonder if asyncio and dma support would be a good combination
you could await until the transfer finishes
hidden rain
Ok, so perhaps a new DMA Module which allow an interface to ADC, SPI, etc would be in order. The CPy asyncio will still not allow the data rates i'm reaching for...
stuck elbow
I experimented with non-blocking spi for displays, and there was almost no speedup, most of the time is spent calculating the pixel data in displayio
blissful pollen
I experimented with non-blocking spi for displays, and there was almost no speed...
Im glad I just read that, I was wondering about trying that out myself to speed displays up
hidden rain
A one-shot ADC thru DMA read to 1000 byte buffer would provide 2 microsecond sampling for 2 milisecond, which could be blocking for that time.
stuck elbow
Im glad I just read that, I was wondering about trying that out myself to speed ...
I might have done it wrong, I admit
slender iron
blissful pollen
I might have done it wrong, I admit
I was looking at it from the point of trying to speed two independent displays up but didn’t get too deep myself
stuck elbow
I was looking at it from the point of trying to speed two independent displays u...
I basically just moved the while loop that waits for the transfer to end from the end of the function to the beginning of it
so it will return, but block when you call it again
blissful pollen
Makes sense. I haven’t had a chance but also was going to look at what the max rate I can even expect from SPI is. It may not even really matter
stuck elbow
if you look at the signal with a logic analyzer, there are long gaps between bursts of transmissions
stuck elbow
one way to maybe speed it up is to make displayio run on the second core on the rp2040 and esp32-xx platforms
but then you risk it will interfere with wifi
blissful pollen
Interesting idea. I didn't look too much at where the real bottle neck is, the write to SPI or calculation of what to write.
I did learn calling display.refresh() twice (by accident) on the same display per update loop slows things down 🙂 oops
stuck elbow
well, it would only makes sense with autorefresh, I think
unless display.refresh() was non-blocking
blissful pollen
If I get a chance I should hook up an analyzer just to see how busy the SPI bus is to see what is slowly me down. Good ideas!
hidden rain
@slender iron For now it will be a new module, analoginfast. It can be renamed or generalized if it makes sense to do that later. Unless you prefer a different name. For the time being, it will be only for the Pico, we can extend it for other platforms later if useful. Please SHOUT if not acceptable.Thanks -lee
tulip sleet
<@252717193496756235> For now it will be a new module, analoginfast. It can be r...
I think we would prefer that it be made part of analogio.AnalogIn.
you'll create an AnalogIn object, and this will be one new call that does a bulk multiple read into a buffer
stuck elbow
I think we would prefer that it be made part of `analogio.AnalogIn`.
@slender iron just suggested a separate module
hidden rain
@tulip sleet: Yes I know that what you had mentioned. Then tannewt chipped in. I am happy either way. The actual code is likely to be taken from dma_capture from the raspberrypi against the pico.
tulip sleet
hmm, well, I would kind of prefer it in analogio.AnalogIn if the implementation is relatively small. It can throw NotImplementedError on ports where it's not done
I can talk to Scott about it. It does not matter too much for now; the bulk of the implementation is not related to where it is
hidden rain
Again Im easy either way. analogio would be much easier.
tulip sleet
stuff like pin and ADC setup can be shared, which is one reason I'd prefer that
it could be shared with two modules, but would require refactoring
I think it would be less work to start with as part of AnalogIn. I am not suggesting a separate class.
hidden rain
Perhaps I should get it working under analogio first then move to new module if you guys decide its better later.
Sorry for all the confusion
tulip sleet
if that's ok with @slender iron, that's what I would like, and I think it will be faster for you
it would just be probably one new API call in AnalogIn
hidden rain
ok... on my private build for now with analogio ..until it works
We already have a analogio.AnalogIn and analogio.AnalogOut. The original question was if i merely extend the current API to have optional arguments or create a new APi like AnalogFastIn.
tulip sleet
I think you could do AnalogIn.read_multiple(buffer, length, rate) or something like that. It will fill the buffer with length samples, clocking at the given rate.
if it is not blocking, then there can be a .read_multiple_done attribute or something like that. I am not wedded to that name
or .read_multiple_remaining. depends on what you can know
hidden rain
I like that too.. so it is!
tulip sleet
i think part of Scott's point is that we don't want to have classes appear/disappear from modules, but instead have separate modules. But in this case I think there's a convenient existing place for this functionality in an existing class.
jaunty juniper
can it be simulated on other ports ? (with less efficiency and whatnot but trying ?)
slender iron
my main concern is the added code size of adding anything to analogin
manic glacierBOT
slender iron
@vague tide I was thinking you could look at ulab too. It may use USER_C_MODULES
vague tide
Anybody tried to compile CircuitPython with USER_C_MODULES=/path/to/ext/lib ? I´m struggling with integrating usqlite library to Circuitpython. Any hint very welcome. The closest information I found about doing this was this document: https://pi3g.com/2021/12/23/integrating-the-bsec-in-circuitpython-a-work-in-progress/
which also mentions the USER_C_MODULES= argument for the build. But one the Circuitpython I use (7.3.2) it never gets used.
oh ok, my post is already here
slender iron
just replied (to you I think) on the forum
vague tide
no, ulab is tightly integrated with the makefile and -as far as i dug- not using USER_C_MODULES
onyx hinge
@slender iron @tulip sleet on https://github.com/adafruit/circuitpython/pull/6667 the issue of the exception types was raised, but those aren't new or changed in this PR. Is there anything else preventing this PR from being merged?
vague tide
so, nobody here ever tried to use a third party C lib with Circuitpython ? if so, how ?
slender iron
generally new modules are added into the main repo
lee just did an example module here: https://github.com/latkinso42/fibonacci/blob/Main/FibonacciStoryFULL.md
unlike MP, CP has split the Python -> C API code out of individual ports
(the shared-bindings directory)
vague tide
i agree there are several examples of integration of bits of C code with CP. However, I want to integrate usqlite lib, which if perfectly functional with micropython (and easy to add thanks to USER_C_MODULES flag) and I don´t want to touch the code. I just want CP to read its micropython.mk or micropython.cmake and nicely fit it in my build.
also, nothing in this lib is port-specific
so what i need is really where in CP i should include that lib, if the USER_C_MODULES flag is not taken into account (it seems it is not, and it seems it used to be)
slender iron
we don't do much with cmake so the .mk is probably the way to go
there are .mk files in py/ and in each port
vague tide
yes, i saw that. But no generic way to tell the build process to grab those ?
onyx hinge
mkdir modules
git clone --depth=1 https://github.com/spatialdude/usqlite modules/usqlite
make USER_C_MODULES=modules BOARD=adafruit_feather_rp2040 clean
make USER_C_MODULES=modules BOARD=adafruit_feather_rp2040 -j9
```I did this, and it gets errors while building the usqlite files so something is happening.i did the same and nothing happened. and i cannot figure out why
which version are you using ?
onyx hinge
678 | .getiter = usqlite_cursor_getiter,
| ^~~~~~~
```so usqlite will require changes before it can work with CircuitPython. we have changed how mp_obj_type_t is organized in order to reduce flash storage, which is the cause of this message.i´m with 7.3.2
onyx hinge
current/recent main branch
vague tide
yes, usqlite will require changes, i expected that, and i´m ready to tackle these.
so the 8.0.0 ?
or 8. something
onyx hinge
did you clone into modules/usqlite and then specify USER_C_MODULES=modules? One thing that initially tripped me up is that an additional directory level must be introduced.
The branch name I am using is "main".
vague tide
i cloned usqlite in lib directory
onyx hinge
in circuitpython
vague tide
ok, i will give a try with the main branch, and i also stumbled upon that same "additional directory level2 thing too
onyx hinge
before I step away I'll try it with 7.3.x branch as well
We did not knowingly break and then fix USER_C_MODULES
vague tide
i wll dig that deeper and report my progress here (and on the forum)
slender iron
here or an issue is a better spot
the forum can be distracting with folks asking for general support
and issue comments are posted here too
onyx hinge
I get the same results with main and 7.3.x branches of circuitpython
Including User C Module from modules/usqlite
and then compiler errors
vague tide
weird. I spend a whole days struggling with that with no effect.
gilded cradle
@slender iron I got the terminal working and am working on the file stuff. I currently have the hash tag params working, but realized we never really discussed what we wanted to do with regards to the password to access the file listings. Is this something we want to allow in the hash or should we just prompt the user for it?
slender iron
<@252717193496756235> I got the terminal working and am working on the file stuf...
Prompt maybe?
The browser may do it automatically
gilded cradle
Ok cool. That's what I was thinking. Since the code is running outside of the device, it's not automatic.
Hmm, I might be able to use a slightly different approach that may automatically prompt them...
slender iron
the browser should prompt I think. IIRC I saw it happen when doing local dev but calling out to the device
I think it's caused by the HTTP response code we send
gilded cradle
Yeah, I'm going to try an xhttp request and I think that would do it.
slender iron
I saw it with the new api just fine
xhttp was only interesting if wanting progress bar for upload
tulip sleet
<@252717193496756235> <@329766224093249548> on https://github.com/adafruit/circu...
I didn't know they weren't new. I will approve and merge
onyx hinge
@tulip sleet thanks!
orchid basinBOT
jaunty juniper
so, we cannot guarantee being able to measure 1.5 ms on a trinket, right ? monotonic() looses millisecond precision after 1h, monotonic_ns() is not available, and ticks_ms is an int.
tulip sleet
so, we cannot guarantee being able to measure 1.5 ms on a trinket, right ? monot...
that sounds right to me
manic glacierBOT
gilded cradle
@slender iron looking at the web workflow circuitpython code, it doesn't look like there's a way to provide an offset in order to do partial file writes like we do with bluetooth. Would this be something that would be difficult to implement?
jaunty juniper
<@252717193496756235> looking at the web workflow circuitpython code, it doesn't...
he's streaming, come ask in the live channel 😉
gilded cradle
he's streaming, come ask in the live channel 😉
Ok, thanks.
edgy flax
Hey - anyone else out there running pre-commit / pylint/ black on windows? Willing to offer an opinion on why I can’t get it to work? Failing with error in cmd window, and failing silently in git bash shell…
manic glacierBOT
manic glacierBOT
orchid basinBOT
manic glacierBOT
orchid basinBOT
lone axle
I've managed to make a new actions task that outputs sizes of the mpy file and strings within it. https://github.com/FoamyGuy/Adafruit_CircuitPython_SI1145/runs/7705514452?check_suite_focus=true#step:14:12
onyx hinge
Neat!
tulip sleet
@onyx hinge have you looked at using asyncio.Event for keypad etc? I was thinking (a while ago) that it's a general mechanism that could be used to asyncify things without having to add async functionality to builtin functions (which is hard to do anyway)
onyx hinge
@tulip sleet not yet. This is using the extension ThreadSafeFlag instead. btw did you make public the neopixel timing spreadsheet? over on qmk someone's dealing with an unusual neopixel like chip that I couldn't find datasheet for, and thought it might be helpful .. but maybe your document was not self-explanatory enough to be worth releasing.
tulip sleet
the arrival of a keypad Event could set the flag in asyncio.Event, and something can await that, so no async operations have to be added to keypad itself
timings I arrived at are in https://github.com/adafruit/circuitpython/pull/6312/files
onyx hinge
I think there's a step there that I don't follow. If keypad doesn't know anything about async, how does something happen as soon as the event arrives?
tulip sleet
keypad is scanning in the background. If a keypad event is queued, an Event flag that is passed into keypad is set when the keypad event is queued. User code can await that Event
onyx hinge
so all the keypad (or queue) does is call an optional callback function?
tulip sleet
that's the general thing, but I was just going to pass asyncio.Event that would get set. I think there is _asyncio.Event??
so it could be native
onyx hinge
I don't think there's a native event object
tulip sleet
ok, no, I see that after looking
but an asyncio.Event is not an interrupt anyway. It's just something that's checked in the asyncio scheduler loop
so maybe it could even monitor just a shared boolean
onyx hinge
I'm also not sure about calling back into arbitrary code during a background task. We do do this when folks use the old python-coded sdcard code.
onyx hinge
so maybe it could even monitor just a shared boolean
This is almost exactly what ThreadSafeFlag does
tulip sleet
it would be nice if it were related to the CPython features of asyncio
ThreadSafeFlag is not native either
onyx hinge
no, it's not native, but what it does in set() is extremely limited
tulip sleet
We can use an asyncio task that decides to run when there is an available keypad.Event. That could be done with pure polling right now that just checks the __bool__ value of the EventQueue
onyx hinge
I am assuming that while we don't have threads, the semantics of the background task and a micropython thread are similar enough (e.g., only called between execution of bytecodes) that I can just not worry about the details
tulip sleet
that's what my examples in the asyncio guide do already, if I remember right
onyx hinge
but then you have to decide how frequently to sleep between checks. await asyncio.sleep(0) doesn't seem ideal, but what is?
tulip sleet
it's just yielding, so something else can run. If there's nothing else to run, it will check again immediately. Are you wanting to light-sleep instead to save power?
onyx hinge
It's more about how much it takes away from other computational tasks, but I guess that's not sensible; a computational task won't await anything so this await sleep will not return
I'm totally fine with this change not going in, of course. I wanted to understand a bit more about how core code could interact with asyncio and it's a bit weird
tulip sleet
There is an open issue in MicroPython to provide some kind of native Event support, I think, for various specific native things. They were wanting to do something like select() on stream, for example. I will find the issue.
onyx hinge
while I've got your attention ... This code frequently crashes (USB stops responding) rather than finishing: ```py
SPDX-FileCopyrightText: 2022 Dan Halbert for Adafruit Industries
SPDX-License-Identifier: MIT
import asyncio
import board
import digitalio
async def blink(pin, interval, count): # Don't forget the async!
with digitalio.DigitalInOut(pin) as led:
led.switch_to_output(value=False)
for _ in range(count):
led.value = True
await asyncio.sleep(interval) # Don't forget the await!
led.value = False
await asyncio.sleep(interval) # Don't forget the await!
async def crash():
1/0
async def main(): # Don't forget the async!
led_task = asyncio.create_task(blink(board.LED, 0.25, 10))
crash_task = asyncio.create_task(crash())
await asyncio.gather(led_task, crash_task) # Don't forget the await!
print("done")
asyncio.run(main())
It never prints `done`.code.py output:
Traceback (most recent call last):
File "/lib/asyncio/core.py", line 214, in run_until_complete
File "code.py", line 19, in crash
ZeroDivisionError: division by zero
[tio 11:47:17] Disconnected
should I file a bug, I suppose?
onyx hinge
eh, maybe <https://github.com/micropython/micropython/issues/2664>
no activity for 4 years doesn't bode well
tulip sleet
gather does not catch exceptions itself
without return_exceptions=True, not sure that's implemented in MicroPyton
exception handling with asyncio is a mess, hence TaskGroups in trio and Python 3.11. Gather does not do exactly what people want
I really thought there's a more recent dicussion of what they want Event to do. Maybe it's just some TODO's somewhere in the code
onyx hinge
I don't know what "does not catch exceptions" means, but at worst shouldn't USB stay connected and I can ctrl-c it if necessary?
tulip sleet
i agree it is a bug, I did not test the exception handling. If you put the gather inside a try-except, does it catch it?
onyx hinge
no, await gather seems never to return somehow
try:
await asyncio.gather(led_task, crash_task) # Don't forget the await!
except Exception as e:
print("some exception occurred, ignoring")
@onyx hinge you could try gather(..., return_exceptions=True)
it is implemented in the library. There are some fixes to the library and the impl upstream which are not incorporated yet; not sure they have anything to do with this
onyx hinge
accepted but no difference
tulip sleet
@onyx hinge you could try disabling _asyncio and see if it still fails when the non-native impl is used
easy to do by editing the library instead of recompiling
tulip sleet
perhaps we should modify the library to not use _asyncio right now
onyx hinge
no fixes in micropython since 1.18 look related
tulip sleet
it could be how we do C scheduling of stuff. They way they do it is pretty different
or it could be due to the way we do async/await, which is a little different (we implement __await__ (courtesy WarriorOfWire) and they don't)
was the neopixel info useful?
onyx hinge
was the neopixel info useful?
I passed it on, too soon to say
the circuitpython build of micropython-coverage's async doesn't work right. It doesn't alternate between two tasks at all
I had to slightly massage circuitpython_ticks for it to work, maybe I got that wrong
using non-native for Task, TaskQueue makes micropython-coverage work as expected. I wonder if there is some mismatch between what we need from Task and TaskQueue, and what our micropython-originated C code is doing.
I feel like I learned a lot so that's good 🙂 now I should do something else with my day that is clearly NOT work 😜
tulip sleet
i am glad you spent time on it and opened the issues, which were kind of on the back burner but I had forgotten about, esp the native vs non-native. I had a confusing time testing that stuff
onyx hinge
You're welcome!
manic glacierBOT
solar crypt
Not sure if I can ask this here but it seemed like the most appropriate channel. What's the status of BLE support on the ESP32-S3? I tried finding a PR but wasn't able to locate one.
jaunty juniper
Not sure if I can ask this here but it seemed like the most appropriate channel....
I'd say: stuck https://github.com/adafruit/circuitpython/issues/5926
solar crypt
Thanks.
nova haven
hey! I'm trying to compile circuitpython for BOARD=adafruit_macropad_rp2040 but I'm getting errors related to rom_hword_as_ptr function like this:
sdk/src/rp2_common/pico_bootrom/include/pico/bootrom.h:129:66: error: array subscript 0 is outside array bounds of 'uint16_t[0]' {aka 'short unsigned int[]'} [-Werror=array-bounds]
129 | rom_table_lookup_fn rom_table_lookup = (rom_table_lookup_fn) rom_hword_as_ptr(0x18);
I've tried building on main and and 7.3.x branch but the result is the same.
Can someone help me to fix it? Maybe I need some specific version of gcc?
$ arm-none-eabi-gcc -v
Using built-in specs.
COLLECT_GCC=arm-none-eabi-gcc
COLLECT_LTO_WRAPPER=/usr/lib/gcc/arm-none-eabi/12.1.0/lto-wrapper
Target: arm-none-eabi
Configured with: /build/arm-none-eabi-gcc/src/gcc-12.1.0/configure --target=arm-none-eabi --prefix=/usr --with-sysroot=/usr/arm-none-eabi --with-native-system-header-dir=/include --libexecdir=/usr/lib --enable-languages=c,c++ --enable-plugins --disable-decimal-float --disable-libffi --disable-libgomp --disable-libmudflap --disable-libquadmath --disable-libssp --disable-libstdcxx-pch --disable-nls --disable-shared --disable-threads --disable-tls --with-gnu-as --with-gnu-ld --with-system-zlib --with-newlib --with-headers=/usr/arm-none-eabi/include --with-python-dir=share/gcc-arm-none-eabi --with-gmp --with-mpfr --with-mpc --with-isl --with-libelf --enable-gnu-indirect-function --with-host-libstdcxx='-static-libgcc -Wl,-Bstatic,-lstdc++,-Bdynamic -lm' --with-pkgversion='Arch Repository' --with-bugurl=https://bugs.archlinux.org/ --with-multilib-list=rmprofile
Thread model: single
Supported LTO compression algorithms: zlib zstd
gcc version 12.1.0 (Arch Repository)
I just tried another rp2040 board and it failed for the same reason
Is there a docker image that I can use to compile circuitpython? I have tried finding if github actions have one but it seems like no
jaunty juniper
I haven't seen that error...
to start did you follow the guide ? there's a few steps to get the right environment and things, including the version of gcc https://learn.adafruit.com/building-circuitpython
we are using gcc-arm-none-eabi-10.3-2021.07 I believe
or the github workflow seems to reference 10-2020-q4 actually
nova haven
I haven't seen that error...
to start did you follow the guide ? there's a few s...
thanks, I was using whatever arm-none-eabi-gcc I had from AUR
downloaded the latest manually and it worked
onyx hinge
hey! I'm trying to compile circuitpython for `BOARD=adafruit_macropad_rp2040` bu...
It looks like this is related to a gcc regression, see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99578 -- You could try adding -Wno-array-bounds to CFLAGS.
hazy walrus
@tulip sleet Finally coming around to working on async requests. I've noticed that the .tox directory in git is not ignored. It looks to be an auto-generated directory, is there a reason why it isn't?
Also thank you again so much for the base to work on. This week I should get some testing done on my board with Sync and Async requests!
Also if you have no idea and there is a pull request I'll let whoever is the code czar make the decision 🙂
tulip sleet
<@329766224093249548> Finally coming around to working on async requests. I've n...
Is there a .tox dir in the repo, or in your environment? You can add .tox to your personal .gitignore. If it's already in the public repo, delete it
hazy walrus
Learn something new every day! Will do
tidal kiln
@lone axle was this intentional?
https://github.com/adafruit/Adafruit_CircuitPython_SI1145/commit/1f73410d5104db0e921b1abebaff22e742e3e58e
lone axle
<@382939733107408897> was this intentional?
<https://github.com/adafruit/Adafrui...
No, definitely not
I was pushing to my fork to test some changes to actions. I did end up in a weird state with my local repo, but did not intend to, or realize that anything pushed to remote
I am not certain of the proper way to revert back to the commit from before that, but that is what will need to happen I think. I am a bit scared to try pushing a fix for it in fear of messing it up further
okay, github desktop application makes the revert very simple. I have reverted and pushed. But definitely happy to have anyone else double check to ensure it's back to normal.
onyx hinge
https://github.com/micropython/micropython/issues/8440#issuecomment-1207485615 hum does this work in circuitpython now? My testing did cover i2c bus scan using the stemma qt port and I think the pins are shared. but I don't know that anyone but me has tried it. (I checked the box on https://github.com/adafruit/circuitpython/pull/6571) -- just reminded of that micropython issue due to fresh activity today.
onyx hinge
@tulip sleet someone over on qmk made a massive list of RGB LED timings https://cdn.discordapp.com/attachments/867530715192885269/1005740259000205312/WS2812b_and_clones.ods
if you believe datasheet numbers of course
devout jolt
Is there a reason why MICROPY_REPL_EMACS_KEYS is not defined for circuitpython/shared/readline/readline.c? Is it just so Ctrl-B will work? I would really like Ctrl-D to delete character and Ctrl-K to kill the line https://github.com/adafruit/circuitpython/blob/main/shared/readline/readline.c#L164
stuck elbow
ctrl-d restarts the board
devout jolt
not mid-line it doesn't, Ctrl-D only restarts the board if it's first char on empty REPL
tame creek
yeah, a traditional Unix terminal device only recognizes ctrl-D as end-of-file at the beginning of a line (if in "cooked" or "line" mode)
devout jolt
(there's an explicit callout in readline.c for when Ctrl-D is midline)
tulip sleet
Is there a reason why `MICROPY_REPL_EMACS_KEYS` is not defined for `circuitpytho...
You could open an issue. People would probably not find it confusing. It does make builds larger, but not by a lot. We could leave it off for the smallest builds.
devout jolt
You could open an issue. People would probably not find it confusing. It does ma...
Thanks. I'll look more into the history and see if there's any obvious "This is why we don't do that"
tulip sleet
I think it was just to save space. I was the only emacs user I knew of early on; no one else requested it.
jaunty juniper
what I would like is support for whatever Terminal does to move the cursor by one word when I do alt-left-arrow, is that in there ?
ah there's ctrl-a for start of line, not practical with my tio bindings but probably good enough
devout jolt
what I would like is support for whatever Terminal does to move the cursor by on...
on emacs-enabled terminals, moving foreward/backward by word is Alt-F / Alt-B. It doesn't look like circuitpython readline supports that
jaunty juniper
oh wait ctrl-a is already enabled ?
devout jolt
Ctrl-E & Ctrl-E work yes
jaunty juniper
well, you learn something every day
stuck elbow
maybe there is an opportunity to save further space on small builds there
slender iron
thanks, I was using whatever arm-none-eabi-gcc I had from AUR
downloaded the lat...
I use arch and haven't had a problem
nova haven
I use arch and haven't had a problem
I installed the latest arm-none-eabi from aur and it didn't work. Maybe I needed a different package
slender iron
I installed the latest arm-none-eabi from aur and it didn't work. Maybe I needed...
I'm using arm-none-eabi-gcc it is version 12.1
ah, I get the errors too
I probably fixed it last time I worked on rp2040
and by "fixed" I mean disabled that warning
manic glacierBOT
lone sandal
sick creek
for web workflow could be something nice with WASI there is bytecode alliance have this WASM micro runtime and wasmer have webassembly package runtime what might work some improvement so could be intereting with circuitpython way on that
onyx hinge
https://github.com/urfdvw/cpy-ol-console huh it looks like this not very far along yet but from the part of the description that's in English it sounds interesting. the github user also contributed to the Chinese translation of circuitpython this week, which was why I was looking at their github.
random junco
https://github.com/urfdvw/cpy-ol-console huh it looks like this not very far alo...
River Wang follows me on Twitter and tweets about it every once in a while. River was a teacher and was building it for that, but is now a software developer.
It's pretty neat
onyx hinge
aha! I didn't recognize them by their github name.
proven garnet
@idle owl but also generally since it's pretty major: I think we're good to flip the libraries to pyproject.toml!
idle owl
<@330227457296957440> but also generally since it's pretty major: I think we're ...
Simply wanted you to be around for it.
proven garnet
I have everything ready and can start tonight, I'll bring it up in the meeting, but I won't be able to listen in live today (at least for the first half)
I guess the only thing that might help is a pause on merges while it's happening, which is why I'd do it in the evening EST.
idle owl
I have everything ready and can start tonight, I'll bring it up in the meeting, ...
Fair enough.
Sounds good.
idle owl
@proven garnet We'll mention the merges thing In The Weeds so folks know. Please come up with a timeframe.
proven garnet
Will do!
onyx hinge
@tulip sleet I don't know exactly what's going on yet, and it's running code that is waiting to be merged, but on this esp32 board (espressif-esp32-eye) I'm seeing stuff that smells an awful lot like memory corruption. Everything kinda works but I get weird errors in which arguments have impossible types or values. In one case it even complains I've called a function with the same kwarg specified twice but it's in the source only once. It involves the camera too so the problem COULD be related to that. .e.g., Traceback (most recent call last): File "<stdin>", line 147, in <module> File "/lib/adafruit_io/adafruit_io.py", line 446, in publish File "/lib/adafruit_minimqtt/adafruit_minimqtt.py", line 596, in publish TypeError: unsupported types for __le__: 'int', 'type' but there's no way that argument could be a type. and it went away when I printed the value of the function parameter. Have you seen any weirdness like that with esp32?
is weird
tulip sleet
I haven't, but this does seem like a memory problem. If you simulate camera use, with and without the reserved buffer, do you see the same thing?
onyx hinge
no, I didn't try faking the camera. It's an obvious thing to do though.
sick creek
earlier I was refering this WAMR as it might able to use those tips what foamguy have showed https://github.com/bytecodealliance/wasm-micro-runtime on the web workflow
tidal kiln
whats CPs current state of f-string support?
idle owl
whats CPs current state of f-string support?
On most boards. Some boards don't support it. AFAIK.
tidal kiln
Adafruit CircuitPython 7.3.2 on 2022-07-20; Adafruit Trinket M0 with samd21e18
>>> foo = "world"
>>> print(f"hello {foo}")
hello world
ah neat. even workin on a wee trinket. 🙂
do we have any general guidance on usage? like using in libs = OK?
idle owl
do we have any general guidance on usage? like using in libs = OK?
I would ping Dan. Maybe it's on all boards now. He would know.
tidal kiln
@tulip sleet thoughts on f-string usage? (see above). also asking because of this PR for more context:
https://github.com/adafruit/Adafruit_CircuitPython_MAX31856/pull/20
proven garnet
On most boards. Some boards don't support it. AFAIK.
I believe @tulip sleet flipped the switch on it
So they all should have it, even the smaller boards
tulip sleet
that's correct, f-strings everywhere now
idle owl
f-strings for everyone! 🎉
slender iron
I don't think our f-string support is complete though
I think folks have found things that work in CPython but not CP
ornate breach
Has anyone run circuitpython on an fpga soft core at all?
idle owl
@gilded cradle If you're comfortable, can you approve this PR? I'd like to get it live. https://github.com/adafruit/circuitpython-org/pull/1014 (It's totally ok if you're not up for it.)
gilded cradle
<@478616919721771019> If you're comfortable, can you approve this PR? I'd like t...
I'll look in a bit
ornate breach
Gotcha, that’s pretty cool. I might check it out at some point when I have a larger fpga that can run litex
idle owl
I'll look in a bit
Thank you! Not sure if Justin's out today, so I thought I would ask.
idle owl
Thank you Melissa!
gilded cradle
You're welcome. It was a simple change.
idle owl
Yep 🙂
gilded cradle
Plus my head has been in the web stuff space for a while, which makes things easy.
idle owl
Bonus!
onyx hinge
<@&356864093652516868> Hey all! It's a scant 18 minutes until the weekly meeting. Please add your notes to the notes document https://docs.google.com/document/d/1Zm1TPFm0mBhFvzQWCashMrcrBoR_d2SQMGtBrvQ7oSE/edit?usp=sharing and if you can, join us in the text channel.
solar whale
lurking today -- may have to leave at any time... Thanks to all for web workflow -- it is working great!
modern wing
Happily lurking 🙂
solar whale
did it work?
blissful pollen
The "just one more run" followed by "OMG what time is it" moments
idle owl
turbid radish
👍
slender iron
I think I can do the core
errant grail
<insert appropriate space analogy here>
turbid radish
Getting married at Kennedy Space Center Visitor's Complex was out of this world
mental nexus
Still time to enter the HACKtablet giveaway: https://gist.github.com/FoamyGuy/642358cf5c8193b2b4b4704d10f021c6
errant grail
Wood, plastic, and luck are the key materials for the tiny consoles.
tidal kiln
@mental nexus break beam sensor wouldn't work?
mental nexus
<@366059101203202052> break beam sensor wouldn't work?
I’m looking to measure the ball position on the lane. It’s to help train folks to pay attention to where they are bowling vs the arrows.
For new bowlers they pay more attention to the pins, want to give feedback about where they are bowling on the lane.
blissful pollen
Could you use the break beam to trigger the distance sensor measurement?
mental nexus
That’s probably possible, but would require something on both sides of the lane. At best case it can just sit on one side of the gutter.
thorny jay
Just, wondering if web workflow should be multi-lingual, and how... If string (web page) are not normal string to be translated. Maybe that is for in the weeds? But there is not urgency, it is mostly curiosity.
mental nexus
End goal is to cantilever an LED strip over the lane to highlight where the ball went.
proven garnet
Thanks! Don't know why the chat was so hard to find!
slender iron
Just, wondering if web workflow should be multi-lingual, and how... If string (w...
I would like it to be. I'm wasn't able to find a common way to have html translations either
it would be nice to have it work through weblate
proven garnet
Happy to do so if there's interest!
@onyx hinge I think I have a PR waiting for the cookiecutter
ember iris
Will you need a couple of folks on standby to help if something goes awry?
but it sounds good to me
proven garnet
I think I should be good solo
solar whale
What could go wrong? 😉
proven garnet
I think anything awful will be limited to a few libraries, not on the massive scale
ember iris
cool, I should be ping-able after 10pm eastern just in case. I don't know much about it either but just in case
proven garnet
Oh, looks like I do need to PR the changes to the cookiecutter, I can get that done today
solar whale
Does the change impact using pip3 install . to install a local version?
idle owl
@proven garnet OK, please make sure cookiecutter is using pyproject.toml. And, consider a guide, we'll see whether it is approved. Thanks!
errant grail
Thank you!
mental nexus
Thanks all!👋
modern wing
Thanks!
random junco
thanks all
proven garnet
Thanks!
blissful pollen
For anyone working on the core with tannewt off for a while and that he does a lot of reviews, feel free to tag me on any reviews you need help on. I do my best to check it but dont always remember.
onyx hinge
I did NOT do well at taking timecodes today
tulip sleet
I did NOT do well at taking timecodes today
we have CPy timecode gadgets if you want one
solar whale
@proven garnet Does the change to using pyproject.toml impact using pip3 install . to install a local version of a library?
tulip sleet
uses a CPX; press reset when starting record. Press button B for each timestamp
idle owl
Or I have code and 3D print files for a Feather using a NeoKey FeatherWing. Same concept though.
onyx hinge
@tulip sleet I have a hot-key I just forgot to use it
tulip sleet
<@329766224093249548> I have a hot-key I just forgot to use it
me too
onyx hinge
Here is the notes document for next Monday’s CircuitPython Weekly meeting. It is at the normal time of 11am Pacific / 2pm Eastern here on Discord. Everyone is encouraged to attend! Please add your hug reports and status updates even if you’ll be attending the meeting - it’s super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and we’ll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/1AQqEMgemmn7GiPWfkrGbyr3ViaGPSAtYIwdclGa3XPE/edit?usp=sharing
solar whale
@tulip sleet should web workflow be working for the feather huzah32?
tulip sleet
we have not tested wifi on huzzah32 yet, so I don't know if it will work
tulip sleet
if you connect to the repl, does something interesting show up in the title bar of the terminal?
that's where the status bar goes when connected serially (it uses escape codes to change the title string)
tulip sleet
i will try it
solar whale
just rebuilt current main - mine was a day old -- still same
Cool! standard wifi_test works on the huzzah32!
solar whale
Adafruit CircuitPython 8.0.0-alpha.1-104-gf1826b054 on 2022-08-08; Adafruit Feather HUZZAH32 with ESP32
>>>
>>> import wifi_test
<Network> MYSSID -77 6
<Network> MYSSID -60 11
None
ip 10.0.0.101
8.8.4.4
ping 0.057
200
circuitpython open issues 573
circuitpython stars 3118
200
This is a test of Adafruit WiFi!
If you can read this, its working :)
200
{'url': 'https://httpbin.org/get', 'headers': {'User-Agent': 'Adafruit CircuitPython', 'Host': 'httpbin.org', 'X-Amzn-Trace-Id': 'Root=1-62f16177-0413b9a44654517004c9cd30'}, 'args': {}, 'origin': 'redacted'}
200
{'timezone': 'America/New_York', 'utc_datetime': '2022-08-08T19:18:16.125789+00:00', 'raw_offset': -18000, 'client_ip': '71.232.151.158', 'dst_from': '2022-03-13T07:00:00+00:00', 'unixtime': 1659986296, 'utc_offset': '-04:00', 'datetime': '2022-08-08T15:18:16.125789-04:00', 'week_number': 32, 'abbreviation': 'EDT', 'day_of_year': 220, 'day_of_week': 1, 'dst': True, 'dst_offset': 3600, 'dst_until': '2022-11-06T06:00:00+00:00'}
done
>>>
however I did discover the stuff to add to the core so that I can define a small wrapper class and then await keyboard events
class AsyncEventQueue:
def __init__(self, events):
self._events = events
async def __await__(self):
yield asyncio.core._io_queue.queue_read(self._events)
return self._events.get()
def __enter__(self):
return self
def __exit__(self, exc_type, exc_value, traceback):
pass
async def key_task():
print("waiting for keypresses")
with keypad.KeyMatrix([board.D4], [board.D5]) as keys, AsyncEventQueue(keys.events) as ev:
while True:
print(await ev)
ctrl-c doesn't break out of the async, that's curious
as soon as you press the button it does
solar whale
Who would have imagined… an NTP synced clock on a Feather Huzzah 32 running CP. 🎉
stuck elbow
is that the new lcd wing by Joey?
solar whale
Yes -- very nice!
I am looking forward to trying some low-power projects with it. Just started playing with it.
stuck elbow
I made a shield with the same chip, but still need to program it
solar whale
@tulip sleet -- this would explain why it is not working for me https://github.com/adafruit/circuitpython/blob/main/ports/espressif/boards/adafruit_feather_huzzah32/mpconfigboard.mk#L14 I can try enabling it ...
yay! `
Adafruit CircuitPython 8.0.0-alpha.1-104-gf1826b054-dirty on 2022-08-08; Adafruit Feather HUZZAH32 with ESP32
import wifi
wifi.radio.ipv4_address
10.0.0.101`
just made these changes ```jerryneedell@Mac-mini espressif % git diff
diff --git a/ports/espressif/boards/adafruit_feather_huzzah32/mpconfigboard.mk b/ports/espressif/boards/adafruit_feather_huzzah32/mpconfigboard.mk
index 2ed08856f..620f6fafc 100644
--- a/ports/espressif/boards/adafruit_feather_huzzah32/mpconfigboard.mk
+++ b/ports/espressif/boards/adafruit_feather_huzzah32/mpconfigboard.mk
@@ -10,8 +10,8 @@ LONGINT_IMPL = MPZ
so increase it to 32.
CFLAGS += -DCFG_TUD_TASK_QUEUE_SZ=32
-CIRCUITPY_STATUS_BAR = 0
-CIRCUITPY_WEB_WORKFLOW = 0
+CIRCUITPY_STATUS_BAR = 1
+CIRCUITPY_WEB_WORKFLOW = 1
CIRCUITPY_ESP_FLASH_MODE = dio
CIRCUITPY_ESP_FLASH_FREQ = 40m
jerryneedell@Mac-mini espressif %
Having some issues with serial terminal -- file browser seems ok
nevermind --- just needed a RESET -- serial terminal working now. -- sort of ...
solar whale
It was explictly disabled for huzzah32
tulip sleet
(was in a meeting)
solar whale
it is on by default for esp32
tulip sleet
I will submit a PR
solar whale
It seem to be working OK with some issues getting into and out of serial terminal -- seems to cause some problems -- still experimenting.
solar whale
I will submit a PR
Thanks -- I thought there might other issues you would be dealing with so easiest for you to turn this on when you are ready. It is great the have CP working on the huzzah32. Thanks!
now to port it to the Huzzah32 breakout!
And I can dust off many old boards...
slender iron
thanks for testing @solar whale
solar whale
No problem -- Glad to help!
After so many years of saying ..sorry no esp32 support ... this will be fun for a lot of people!
random junco
Does the web workflow work on a PyPortal Titano as it has an ESP32 chipset?
idle owl
@proven garnet While you're doing CP things tonight, can you check on whatever is involved with Adabot producing the CP Libraries report that we use in the CP Weekly, and make sure it's still good to go? It was a rate limit that stopped the report from being generated last night, so it seems unlikely that it will continue, but maybe poke it anyway, if you could. Thanks!
solar whale
Does the web workflow work on a PyPortal Titano as it has an ESP32 chipset?
On the Titano, the ESP32 is not running CP -- it is using the NINA firmware, so I think it not part of the web workflow...yet
random junco
Thanks! I put the latest firmware on it and it wasn't seen, so that makes sense
solar whale
web workflow depends on CIRCUITPY_WIFI https://github.com/adafruit/circuitpython/blob/main/py/circuitpy_mpconfig.mk#L486 so I think it has to be one of the boards with native wifi.
slender iron
we have talked a bit about native ESP32SPI support so web workflow would work. definitely not coming soon
@tulip sleet think it should be @<line number> <filename> <exception name>? or drop exception name
tulip sleet
<@329766224093249548> think it should be `@<line number> <filename> <exception n...
I'd include the exception name -- you could truncate it to a dozen chars or so if you feel it would push everything else off the right end on a TFT, etc.
slender iron
k, I have it included.
we can see how well it works in practice and adjust accordingly
slender iron
I think this will be context for our meeting later: https://github.com/micropython/micropython/pull/8914
proven garnet
<@135415127879974912> While you're doing CP things tonight, can you check on wha...
Chances are it's my doc status checker 😅
I tried to manually slow it down but maybe it's not enough
Is it possible to increase it? It may be worth silencing until we find and API friendly way of doing it, or can have the limit raised
(RTD limit, that is)
Do you have the error somewhere that I can confirm that?
I think I found it, it's actually the existing adabot code, I'll look into it tonight :)
tulip sleet
@slender iron @onyx hinge I thought it was https://github.com/micropython/micropython/issues/9018 that was what jimmo wanted to talk about; maybe both
proven garnet
Alright, pyproject.toml switch is starting now!
manic glacierBOT
proven garnet
We're all set! Everything is upgraded to pyproject.toml! I checked the CI for the builds and I fixed the few issues that popped out.
The only thing left to do is release them all (a test release of one implies that it should go smoothly!)
Please don't hesitate to ping me on GitHub or Discord if anything weird happens as a result, but so far it looks like a painless transition!
ember iris
We're all set! Everything is upgraded to `pyproject.toml`! I checked the CI fo...
Congrats and thanks for the migration!
manic glacierBOT
manic glacierBOT
hmm, not sure why pre-commit is failing here. it passed when I ran it locally.
proven garnet
Oh yeah, added bonus is that libname.__version__ should now work no matter where the library is downloaded from (PyPI was the last hold out), hopefully that helps the folks doing support!
tulip sleet
looking now to see if the chips support SDIO
looks like there is a peripheral but we don't support it yet
spiral elk
I was just looking at starting a board definition for the ESP32-CAM, which has an sdio device onboard
tulip sleet
@onyx hinge is already working on the ESP32-CAM, so you might hold off
spiral elk
Actually looking at his repo it looks like what he's working on isn't the ESP32-CAM board but generalized camera support for espressif.
tulip sleet
I think he may be testing with an ESP32-CAM board, though. @onyx hinge can you comment when you're on?
spiral elk
He seems to be working with ESP32-S3-Eye and Kaluga
I actually don't even know where I put the camera modules for my boards.
onyx hinge
no sdioio support on espressif?
there's not currently sdio support on espressif. I'd love someone to tackle that. An in progress pull request of mine adds support for esp32-eye and esp32s3-eye boards from espressif, as well as adding the esp32-camera library as a circuitpython module.
(and kaluga but there was already a board def for that)
suspect it would be possible to use the (probably slower SPI-based) sdcardio with the sd card, since espressif is so flexible about pin assignments
manic glacierBOT
onyx hinge
and I should fess up that I've been seeing some weird behavior on the esp32 board while the camera is working, I haven't been able to figure out anything 'systematic' yet but it feels like memory corruption of either the stack or the circuitpython heap.
onyx hinge
@slender iron Right now the property names in that PR match the esp32-camera C API. I agree that better documentation is critical, but I am not sure whether to make what I think is a "better name" in the API or stick with the underlying API name. I'll do docstrings first but your feedback about this specific item would be valuable.
slender iron
@onyx hinge I think its fine to leave the names as-is. mainly I want substantive doc strings so one doesn't need to reference the C api docs
new arm toolchain: https://developer.arm.com/downloads/-/arm-gnu-toolchain-downloads
onyx hinge
The API documentation is sparse about the meaning of these flags, it's a puzzle to figure it out.
as are the manufacturer datasheets
slender iron
if its unclear, then omit it
onyx hinge
anyone know what this message from sphinx means, just off-hand? Otherwise I can search about it: WARNING: Inline interpreted text or phrase reference start-string without end-string.
jaunty juniper
I remember seeing it a few times and being baffled by it, don't remember what it really meant ? Maybe a missing empty line in ..source section or similar ?
onyx hinge
yeah it's about a .. seealso::
jaunty juniper
there's strict rules about empty lines and indentation to find the beginning and ending of stuff
jaunty juniper
like it has to be like this ?
<empty line>
.. thing::
<empty line>
the thing
<empty line>
trying it...
it's too bad the turnaround is not a little faster building the core docs
maybe there's a trick I could stand to learn
jaunty juniper
I'd love such a trick
tulip sleet
maybe there's a trick I could stand to learn
do you mean locally?
onyx hinge
@tulip sleet yes
onyx hinge
it seems to take a similar length of time to "make html" as to make a firmware
tulip sleet
and -j<n> doesn't really help
onyx hinge
It was bad syntax on a link, the fact that it was in a .. thing:: block was a red herring.
jaunty juniper
ugh
onyx hinge
I don't write enough .rst for ````... <...>_``` to flow off the fingers. I was missing the trailing \ or something
❗ doesn't seem like the right image to use here but that's not a fight for right now.
onyx hinge
-//| This object is usually used with a camera-specific wrapper library such as `adafruit_ov5640 <https://circuitpython.readthedocs.io/projects/ov5640/en/latest/>_.
+//| This object is usually used with a camera-specific wrapper library such as `adafruit_ov5640 <https://circuitpython.readthedocs.io/projects/ov5640/en/latest/>`_.
``` yup that was it, a missing \`I was gonna say it might have something to do with backticks or number of backticks
onyx hinge
thanks, I appreciated your help brainstorming it
jaunty juniper
I didn't actually think missing one
but I did battle with links for the matrix, so that might have been where I had that message ?
sphinx is so mysterious to me
onyx hinge
@slender iron assuming it comes up green, the esp32-camera PR is ready for you to review again.
trim elm
@proven garnet How's the pyproject.toml stuff going? Would now be a good time for me to do a release sweep?
slender iron
@onyx hinge kk, will look shortly. haven't done my sweep yet
onyx hinge
@slender iron OK, I'll actually click to re-request your review if I see CI has finished green later
proven garnet
<@135415127879974912> How's the pyproject.toml stuff going? Would now be a good ...
Patching a minor mishap for all the pyproject.toml files but should be good in a little bit! I can ping you :)
trim elm
Patching a minor mishap for all the `pyproject.toml` files but should be good in...
Cool! thanks!
slender iron
@tulip sleet 1@filename is fine with me. I wanted the line first though for led matrices
tulip sleet
that sounds good
tulip sleet
i'm not sure, maybe it's harder to read. For instance above, it looks like an "i". Yeah, I like better @ better after all
slender iron
👍 will tweak it now
jaunty juniper
seems reasonable to me, error messages and tools like ag will give: file:line (which you can feed to things like the bbedit command line to open and select that line) but I agree it's better to have the line first
proven garnet
Ha! pyproject.toml fix complete, fixed the issue where dependencies weren't getting added to the built distributions.
Bonus, we're also now building wheels besides source distributions, which should cut down on time from having to build from source distributions when installing from PyPI.
@trim elm you're good to go for releasing everything in the bundle!
trim elm
@proven garnet Awesome! So the thing I'm releasing them for is switching to pyproject.toml, correct?
proven garnet
Yup! They also now have the ability to do libname.__version__ if that's defined for the library when downloaded from PyPI.
trim elm
Oh awesome
proven garnet
But pyproject.toml is the major change.
proven garnet
Seesaw is already released, just a heads up
idle owl
@tulip sleet Are we still having issues with LC709203 and ESP32-S3? The page isn't published in the Feather guide, and I didn't want to add it to the TFT Feather guide if it was still iffy.
tulip sleet
YES
idle owl
Got it.
tulip sleet
there was supposedly a fix but I did not get it to work
idle owl
Good to know. Thank you!
idle owl
<@330227457296957440> mind taking a look at this adabot patch now that setup.py ...
Merged once I fixed the URL 😄
proven garnet
Thanks!!
trim elm
@proven garnet (cc:@idle owl) the libraries are all released
idle owl
Great!
proven garnet
<@135415127879974912> (cc:<@330227457296957440>) the libraries are all released
Running a check on the release.yml runs now!
wraith crow
@slender iron I'm not sure my question regarding the Feather Huzzah build size warrants an issue yet, I haven't even been using clean builds yet. I'll keep an eye on it and if it doesn't start making more sense to me, I'll either query discord or open an issue. Thanks!
slender iron
👍 we have changed relatively recently how we allocate the mp heap so there is definitely room to tweak it
proven garnet
We're golden! All libraries are now published with the changes, thanks @idle owl and @trim elm!
onyx hinge
@slender iron did you deliberately push to jepler/espressif-camera-2 or was it a mistake? oh, was that merging main into the branch, that would make sense.
slender iron
yup, did the merge for you
anyone have cycles to help update: https://learn.adafruit.com/circuitpython-tv-zapper-with-circuit-playground-express/circuitpython-code-2 ?
onyx hinge
"is this something I'd have to have a TV to understand?" <-- I hardly ever get to use that line anymore
thanks for the merge, scott
blissful pollen
Gotta love seeing a TODO in the esp idf code itself
manic glacierBOT
lime trellis
has the topic of dynamically setting pin states before/after deep sleep ever come up for CP development? I know at least the SAMD port resets pins before deep sleep
slender iron
has the topic of dynamically setting pin states before/after deep sleep ever com...
you can probably cheat it through pin alarm
lime trellis
you can probably cheat it through pin alarm
not sure what you mean. wouldn't pin alarm require the pin be deinit first?
sick creek
all the filesystem on web workflow i got dig deeper to WASI filessytem and now also WASI with Wagi
slender iron
not sure what you mean. wouldn't pin alarm require the pin be deinit first?
Ya perhaps. Pin alarm should pull in the opposite direction of the alarm, even when asleep
slender iron
all the filesystem on web workflow i got dig deeper to WASI filessytem and now a...
I don't know what WASI or Wagi are
sick creek
I don't know what WASI or Wagi are
this seems to explain more those 3 https://www.fermyon.com/blog/wasm-wasi-wagi what might be interesting for circuitpython web
slender iron
and how is this related to web workflow?
lime trellis
Ya perhaps. Pin alarm should pull in the opposite direction of the alarm, even w...
interesting thought. I'll investigate.
Am I crazy in thinking this warrants more than a hacky solution? There are lots of cases where you'll want to leaving something either on or off during deep sleep (conditionally)
slender iron
@lime trellis ya, its worth thinking about
jaunty juniper
you can open an issue or see if there is already one
lime trellis
you can open an issue or see if there is already one
yeah will do. Thanks for the encouragement 🙂
manic glacierBOT
sick creek
on web workflow user could run they CPy code on device and see it in webworkflow for example perhaps
manic glacierBOT
I loaded the latest S3 bits, killed my code.py application so we're sitting at the REPL prompt and then started refreshing the file manager. After six refreshes I hit an error but the console message was not one of the usual ones:
I then started hitting refresh again and after another 10 or so refreshes I got another miss-load with the typical `ERR_INVALID_CHUNKED_ENCODING 200...
blissful pollen
I think I have an issue where something is being optimized out. To clear a i2c flag I have to read one of the registers (self->peripheral->hw->clr_rd_req). But I think it may be getting removed cause it doesn't "do" anything. Any idea on how to ensure that read runs?
onyx hinge
yup, did the merge for you
looks that means the pr needs a fresh review approval, as well as finishing its run (gulp)
slender iron
I think I have an issue where something is being optimized out. To clear a i2c f...
hardware registers should be marked as volatile to prevent removal
blissful pollen
Thanks I'll check that. I tried printing the value to ensure it was read and it still didn't clear so.... 🤷🏻♂️
onyx hinge
hmmm the auto-merge happened before the checks completed. I guess I shouldn't touch that knob. Thanks Scott, hope nothing turns out to be wrong
slender iron
np, can fix later
slender iron
Thanks I'll check that. I tried printing the value to ensure it was read and it ...
generally you write a 1 to a register bit to clear
blissful pollen
rp2040 shows it just being read, but worth a shot too.
Plus side I have most of the i2cperipheral working I think, just making sure error conditions don't leave things in a weird state.
slender iron
huh, interesting
blissful pollen
Found out it is clearing other interrupts. I'm just getting into some error state where it is confused if you don't respond quick enough to a read request. Hoping it is something we can detect cause it is a likely error case
manic glacierBOT
Fixes #6723
Emojis are better supported across systems than symbol code points.
ℹ️ 📁 ⬇ 📄
hazy walrus
@tulip sleet Your async code looks like it's working! I've got a little CPython test that I think shows it working with a Mocket and and will load it up with CircuitPython somewhere this week for a real test. It looks like you had merged in an experimental branch with several changes that were different from adafruit's main. I did my best to unwind them to bring it more in line with main. If I were to make a pull request where should it go to?
tulip sleet
<@329766224093249548> Your async code looks like it's working! I've got a little...
you mean in requests? Yes, I was cleaning up exceptions too. The exceptions fix could be done separately.
I got some pushback about whether to include this in the original library or make a separate library. There are some refactoring decisions to make.
Thank you for testing! I did not expect it to necessarily work right away. I assume you fixed some things.
tulip sleet
@onyx hinge the select for EventQueue looks very interesting and could be a model for other stuff. I will look at it more tomw.
onyx hinge
@tulip sleet I'll try to prove it's faster than the Python code, otherwise it's not really super useful
and yes if we want to go that way it shows how to make many core things "nearly awaitable", with a small snippet of Python code. Too bad that Python code has to touch a non-portable part of the asyncio implementation, asyncio.core._io_queue.
tulip sleet
there is some open PR or issue in micropython that talks about how to do select stuff for asyncio for pin waiting and maybe other things. I lost the reference and haven't refound it, but it's what I mentioned yesterday
onyx hinge
maybe it's this discussion? https://github.com/micropython/micropython/issues/2664
g'night
spiral elk
quick question on board def BCP: if the onboard LED isn't broken out to a board pin should you still define board.IOxx or just board.LED?
solar whale
@onyx hinge What boards and cameras are supported by the new camera stuff that you have been working on? Nevermind -- I found waht I was looking for in the PR. I should have looked there first...sorry.
onyx hinge
<@426200547024961539> What boards and cameras are supported by the new camera st...
The work is for Espressif microcontrollers (ESP32, ESP32-S2, and ESP32-S3) and should be for "all cameras that esp32-camera supports" (13 different sensors). My actual testing was with OV2640 and OV7670.
solar whale
The work is for Espressif microcontrollers (ESP32, ESP32-S2, and ESP32-S3) and s...
Thanks. I’ll attempt to try it out.
hazy walrus
you mean in requests? Yes, I was cleaning up exceptions too. The exceptions fix ...
Nope, I don't think I made any functional changes 🤔! I'll get it pushed up at work this morning and get a few more tests in there + a real test on my board at some point. My wife is out of town so I've got time to kill 🙂
As far as inclusion in the library I just ended up having it be a "peer" of the adafruit_requests.py file. I don't have any huge problems with any of it and am happy to help package it any way it needs to if people don't have the time.
Is there a good parallel anywhere in adafruit's librarys where a non-asyncio library was wrapped up and made working with asyncio for a template on what to do?
Since it really just seems like it's copy-pasted code with a few asyncs thrown in here or there it seems like it's just a wrapper that adds support for a language feature so I'd kind of err on keeping it simple? I'm also not a maintainer of anything circuitpython.
Usually cowboy code doesn't work but in your case it seems like it did 🤩
tulip sleet
Nope, I don't think I made any functional changes 🤔! I'll get it pushed up at ...
See https://github.com/adafruit/circuitpython/pull/6712 and https://github.com/adafruit/circuitpython/issues/6736 for some very recent work on this. In the requests library, it would be nice to have less duplicated code between the async and non-async versions, but due to the "infectiousness" of async/await, that was not so easy to do. There's no way to have a little conditional inside that either does an await or doesn't because the whole routine has to be async if you include an await.
We freeze adafruit_requests on some boards, so kind of doubling its size to include the "peer" async version would be problematic in terms of space. At the same time, having it in the same library makes it easier to maintain the duplicated code. It's a bit of a dilemma we have to think about.
I really appreciate your testing, and that it seemed easy to use.
hazy walrus
See <https://github.com/adafruit/circuitpython/pull/6712> and <https://github.co...
Yes, I'm trying to get an idea of where the non async parts are and I think it's all just in the socket code that's implemented per board maybe? It's my first time working with a microcontroller so I'm just sort of shooting in the dark. The test I'm currently trying to get running is opening 3 sockets that are mocked to have some sleep or delay in the received and then make sure all have SENT
The only problem is that it looks like the non async sleep blocks the whole thread which blocks any coroutine running. So not a silver bullet but I think it could still work pretty well.
My big concern would be how is the lower level code working, how many assumptions are being made that there's only one socket open at any time on any given board?
tulip sleet
i don't think we assume only one socket, but the fetches are all blocking right now
onyx hinge
@tulip sleet we had chatted a bit about how my esp32-eye board is acting flaky with impossible exceptions. I have re-organized the example that was having problems and it occurs even when the new camera module is not even used, and when esp-idf reserved memory is set to 0 or non-zero. Doing things that change memory allocations make a different weird problem happen at a different spot, or sometimes just happen to make things work right. I'm thinking I should try the other module of the same type I have.
tulip sleet
<@329766224093249548> we had chatted a bit about how my esp32-eye board is actin...
have you tried similar tests on just a plain ESP32 board (HUZZAH32 or V2?) Does that make sense to try?
I would not be surprised if there were some memory issue
onyx hinge
yes that would also make sense to try, right now I'm just trying to do a basic MQTT publish example now that I've started reducing my code
MQTT has been flaky on me generally, but this is "impossible errors" not "sockets are maybe unreliable errors"
Traceback (most recent call last):
File "<stdin>", line 55, in <module>
TypeError: function got multiple values for argument 'socket_pool'
print("Connecting to Adafruit IO")
mqtt_client = MQTT.MQTT(
broker="io.adafruit.com",
username=aio_username,
password=aio_key,
socket_pool=pool,
ssl_context=ssl.create_default_context(),
)
I'm gonna try and get my build environment up to date and get the ESP32-CAM up and running tomorrow. Would the new camera-code potentially run on it? The board has an OV2640
tulip sleet
MQTT has been flaky on me generally, but this is "impossible errors" not "socket...
I wonder if this is a cache problem; maybe you said that yesterday
onyx hinge
@spiral elk yes, OV2640 is supported by the library
stuck elbow
OV2640 seems like a very nice compromise between price and features
hazy walrus
Is this just a fact of life when looking through the circuitpython code forever?
tulip sleet
Is this just a fact of life when looking through the circuitpython code forever?
you can apply to be part of the https://cs.github.com code search beta
i am in the beta; it's a very nice UI
onyx hinge
Depending on your goals using a local search tool like "git grep", "ag", etc., can give very quick and accurate results for string & regex searches.
tulip sleet
^^ that is what we all do most of the time. I also use emacs tags to wander around definitions and uses
onyx hinge
Code in forks is only searchable if the fork has more stars than the parent repository
FBOFW we are not likely to ever get more stars than micropython
hidden rain
Attempt to Extend analogio. https://github.com/latkinso42/circuitpython/tree/adcdma
Code loaded once, then crashed Pico. After that, the Pico will not connect with that UF2. However it will connect with noraml load UF2.
I've gone over the code multiple times, and recompiled mutiple times. Still will not connect
onyx hinge
@tulip sleet I'm not reproducing the problem on esp32 feather v2
@hidden rain you've built & loaded other circuitpython uf2s yourself?
hidden rain
Yes. Example the Fibonacci code just mentioned in the last newsletter
hidden rain
yes. but again. at this point the UF2 loaded Pico will not connect to my PC thru USB
onyx hinge
if so I recommend booting in safe mode (refer to board docs for how), then rename code.py to something else like nocode.py. Then, to test it, ctrl-c to get to the repl and import nocode to test it.
If your code malfunctions "in certain ways" it will stop performing the activities that are needed to connect on USB or to maintain the connection. When this happens very early incode.py, safe mode is the way to get the device connectable again.
the "certain ways" can be as simple as having an infinite loop that never terminates.
hidden rain
Sure .. I understand. That means going thru safe mode to make the determination, then avoiding the programming error in the C-Code. OK....
onyx hinge
by keeping the test python code in a different filename than "code.py" it's a little bit less frustrating. Another thing I've used is this snippet of code that waits for USB serial to connect & then for you to hit enter to run: ```py
import supervisor
import time
import sys
wait for serial connection
if supervisor.runtime.serial_connected:
print("serial already connected, continuing")
else:
while not supervisor.runtime.serial_connected:
pass
time.sleep(.1)
while supervisor.runtime.serial_bytes_available:
sys.stdin.read(1)
print("press any key", end='')
while not supervisor.runtime.serial_bytes_available:
pass
sys.stdin.read(1)
print()
last note for now, it's possible to set up one pico to debug another pico -- this is called picoprobe and works in conjunction with debugger software on your computer. I've done it on linux but it's been awhile. https://community.element14.com/products/raspberry-pi/b/blog/posts/debugging-the-raspberry-pi-pico-on-windows-10
hazy walrus
Is there a high level overview of how the circuitpython code is laid out?
Do micropython updates get brought into circuitpython?
hidden rain
@onyx hinge Really Cool Stuff! Thanks ! Off to safe mode and or debug...
onyx hinge
@hazy walrus https://learn.adafruit.com/extending-circuitpython/overview and https://github.com/adafruit/circuitpython/blob/main/docs/common_hal.md are some of the docs about how we organize modules and ports. Yes, we merge in micropython from time to time; I think we're based on 1.18 currently (in 8.0 alphas)
onyx hinge
Want to contribute to the CircuitPython core? Learn how we organize ports & importable modules: https://learn.adafruit.com/extending-circuitpython/overview and https://github.com/adafruit/circuitpython/blob/main/docs/common_hal.md
cobalt grail
Not sure how to handle this - Weblate tells me to localize the string "All I2C targets are in use", but there is already a the localized string "All I2C peripherals are in use" for this.
Am I supposed to make a pull request to the code base to change this?
onyx hinge
@cobalt grail Please do! Fixing near-duplicate strings in the source code is appreciated.
So just a note though
I think "target" is the correct new terminology. Let me find you a reference ...
so I would think the old message about "peripherals" should have been removed when we changed
cobalt grail
Well "target" makes more sense to me, but is harder to translate to swedish. 🙂
onyx hinge
You can ask nxp what the correct translation is.
You may want to open an issue instead, I'm not 100% sure which message should be changed to what..
cobalt grail
"Target" translates to "mål" which also means "goal" in swedish.
onyx hinge
because the messages are actually about the specific peripheral/hardware block within the microcontroller, which can be used as a target or as a controller.
cobalt grail
OK, i'll open an issue then.
onyx hinge
I appreciate you noticing things like this! I know I frequently don't pay it adequate attention, as a developer
cobalt grail
I appreciate you noticing things like this! I know I frequently don't pay it ade...
Thanks. I don't even have the code on my machine, as weblate makes localization so streamlined. Even links to the exact code line in the codebase.
slender iron
quick question on board def BCP: if the onboard LED isn't broken out to a board ...
If it is labeled in silkscreen, then I'd add another name besides LED.
slender iron
slender iron
I disabled auto-merge since it is ignoring the CI.
tulip sleet
I disabled auto-merge since it is ignoring the CI.
oh, really, oy
slender iron
looks like the branch protections weren't set
I added some status checks to them but there doesn't seem to be an "all" checkbox
tulip sleet
i noticed that I couldn't set automerge after the fact, which seemed weird
slender iron
yup
spiral elk
If it is labeled in silkscreen, then I'd add another name besides LED.
In the case I'm looking at, silkscreen is only labeled LED so I guess I'll omit the additional name.
manic glacierBOT
slender iron
@wraith crow I've got a fix for the C3 that I'd love you to test when you have a chance
gilded cradle
@slender iron I've been trying to increase the reliability of detecting disconnects for the web workflow stuff. I tried using web sockets because of the ping response option, but ultimately found there's no way to directly send opcodes with the native WebSocket. I was next considering occasionally checking for one of the static files, but was curious if you had any better ideas.
Currently if I unplug the board, it still seems to think it's connected.
tidal kiln
@slender iron back (hopefully) from power outage. looking at pr now.
slender iron
@gilded cradle why not just wait until another action tries to talk to the board?
gilded cradle
I wanted to change the disconnect button, but that is an option.
slender iron
I think its ok to re-fetch the version.json
I think websocket will know it is closed eventually
gilded cradle
Yeah, but not until I plug in the board again
gilded cradle
I think websocket will know it is closed eventually
It's ok. I was just hoping you had maybe already had some simple way in mind.
slender iron
nope, sorry
jaunty juniper
how are you going about having the same tools on cp.org for the web workflow and the BLE workflow ? (And the USB drive)
gilded cradle
I have workflow classes with common functions. You will choose a connection type eventually, but right now I have it so you can pass parameters in the url.
You could check out what I have so far at https://github.com/makermelissa/web-editor/tree/webworkflow, but's it's still a work in progress.
jaunty juniper
so, like a higher level API that abstracts the backend, so I can plug my library manager into it ?
gilded cradle
so, like a higher level API that abstracts the backend, so I can plug my library...
Yeah, I've been abstracting it. I'd like to eventually be able to allow the library manager to work with it.
Or it to work with the library manager, whichever.
jaunty juniper
I want to look at the web-editor repo, but so far I don't understand why I need a token to npm install, that's weird
slender iron
the ble library isn't hosted by npm I think
its installed directly from github IIRC
gilded cradle
I think it was because we were using github to host an npm package. However, we should be able to use jsdelivr like I did here: https://github.com/adafruit/WipperSnapper_Firmware_Uploader/blob/main/js/script.js#L7
But this was from before I learned that.
wraith crow
<@899160869777342464> I've got a fix for the C3 that I'd love you to test when y...
Okay, I'll grab the PR and update it with my results 🙂
slender iron
Okay, I'll grab the PR and update it with my results 🙂
thank you!
