#circuitpython-dev
onyx hinge
Howdy <@&356864093652516868> ! In about 50 minutes it'll be my pleasure to welcome you to our weekly meeting. please add your notes to the document: https://docs.google.com/document/d/1C4XIYb0tcKxlzNpwMI5pGf2OmRlzvSLOz7nbUt_Jo3s/edit?usp=sharing
manic glacierBOT
onyx hinge
@lone axle do you anticipate being able to read the library section of the meeting today?
lone axle
<@382939733107408897> do you anticipate being able to read the library section o...
Yes, I believe I should be able to.
manic glacierBOT
onyx hinge
in "raw repl" on a Adafruit CircuitPython 9.1.0-beta.3-48-g396aaef964-dirty on 2024-06-28; Adafruit Feather ESP32 V2 with ESP32 I can consistently paste a 103 line 7994-byte string via tio 2.6. In "paste mode", even a 100ms "line delay" (-O100) is not enough. A 1ms "char delay" (-o1) is enough.
this is just using a terminal keystroke to paste
wraith crow
The "raw repl" mode doesn't echo characters as you're typing/pasting does it? The echo is what has been causing the problem I think
turbid radish
βΊοΈ
lone axle
Here is an example of the ruff config https://github.com/adafruit/Adafruit_CircuitPython_Requests/pull/173/files#diff-50c86b7ed8ac2cf95bd48334961bf0530cdc77b5a56f852c5c61b89d735fd711R41
onyx hinge
justmobilize's:
ignore = ["PLR2004", "UP030"]```
my project (leapseconddata):
select = ["E", "F", "D", "I", "N", "UP", "YTT", "BLE", "B", "FBT", "A", "COM", "C4", "DTZ", "FA", "ISC", "ICN", "PIE", "PYI", "Q", "RET", "SIM", "TID", "TCH", "ARG", "PTH", "C", "R", "W", "FLY", "RUF", "PL"]
ignore = ["D203", "D213", "D400", "D415", "ISC001"]```
spare jacinth
i havent used it, but i know the ruff TOMLs support some sort of composition
at the very least i saw you can override "default" library-level settings in some subfolder
could perhaps leave the pyproject's as is, and add an extra toml with the enable/disable rules
spare jacinth
$ ruff linter
F Pyflakes
E/W pycodestyle
C90 mccabe
I isort
N pep8-naming
D pydocstyle
UP pyupgrade
YTT flake8-2020
ANN flake8-annotations
ASYNC flake8-async
S flake8-bandit
BLE flake8-blind-except
FBT flake8-boolean-trap
B flake8-bugbear
A flake8-builtins
COM flake8-commas
CPY flake8-copyright
C4 flake8-comprehensions
DTZ flake8-datetimez
T10 flake8-debugger
DJ flake8-django
EM flake8-errmsg
EXE flake8-executable
FA flake8-future-annotations
ISC flake8-implicit-str-concat
ICN flake8-import-conventions
LOG flake8-logging
G flake8-logging-format
INP flake8-no-pep420
PIE flake8-pie
T20 flake8-print
PYI flake8-pyi
PT flake8-pytest-style
Q flake8-quotes
RSE flake8-raise
RET flake8-return
SLF flake8-self
SLOT flake8-slots
SIM flake8-simplify
TID flake8-tidy-imports
TCH flake8-type-checking
INT flake8-gettext
ARG flake8-unused-arguments
PTH flake8-use-pathlib
TD flake8-todos
FIX flake8-fixme
ERA eradicate
PD pandas-vet
PGH pygrep-hooks
PL Pylint
TRY tryceratops
FLY flynt
NPY NumPy-specific rules
AIR Airflow
PERF Perflint
FURB refurb
RUF Ruff-specific rules
slender iron
wraith crow
Thanks for hosting Jeff π
candid sun
thanks Jeff!
lone axle
Thanks for hosting Jeff. Have a great week everyone!
random junco
thanks Jeff!
manic glacierBOT
onyx hinge
https://gist.github.com/jepler/117ceae99f8c9439f86fcdd4e28d3f9b here's what pylint-to-ruff prints (minus comment lines) when I run it in adafruit_circuitpython_requests
so we could just chuck that in ruff.toml I guess
# 4. Ignore `E402` (import violations) in all `__init__.py` files, and in select subdirectories.
[lint.per-file-ignores]
"__init__.py" = ["E402"]
"**/{tests,docs,tools}/*" = ["E402"]```you can also apply different settings by path so we can invoke ruff once, not twice (right now we lint examples differently)
lone axle
A handful of the ruff rules that were output by pylint-to-ruff tool end up outputing these warnings when ruff is run:
warning: Selection `PLC2801` has no effect because preview is not enabled.
warning: Selection `PLE0643` has no effect because preview is not enabled.
warning: Selection `PLE0704` has no effect because preview is not enabled.
warning: Selection `PLE1141` has no effect because preview is not enabled.
warning: Selection `PLR0202` has no effect because preview is not enabled.
warning: Selection `PLR0203` has no effect because preview is not enabled.
...
There are about 20 of them. Would we want to run with "preview" enabled in order to use these? Or if not should we remove them from the config file for now if they will have no effect?
onyx hinge
ugh my audio was clippy during the recording. oh well.
lone axle
I have updated the BME680 PR to move the ruff config into it's own file and bring in those rules that were outputed by pylint-to-ruff. https://github.com/adafruit/Adafruit_CircuitPython_BME680/pull/70 for now I've left all the ones that it warns about having no effect. There were a few other tweaks to the code from running the ruff with these rules in place but not too much.
unrelated to ruff, but it looks like perhaps we're going to face this on lots of libraries (I saw it on a different one earlier today):
Error during sphinx build
/home/runner/work/Adafruit_CircuitPython_BME680/Adafruit_CircuitPython_BME680/adafruit_bme680.py:docstring of adafruit_bme680.Adafruit_BME680_I2C:19:'any' reference target not found: board.I2C
The cause as far as I can tell is useage of tick marks surrounding "board.I2C" in a docstring. By changing them to double tick marks it makes it so that the sphinx build can complete successfully.
onyx hinge
Here is the notes document for next Mondayβs CircuitPython Weekly Meeting. It is at the normal time of 11am Pacific / 2pm US Eastern here on Discord. Add your hug reports and status updates to the document before the meeting. If you are unable to attend but would still like to contribute, feel free to add your notes and weβll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/1cvCxFaJT5hIQ4UmzZm6f_betgOKNhtXJrsA4GYdVeQ4/edit?usp=sharing
onyx hinge
Here is the notes document for next Mondayβs CircuitPython Weekly Meeting. It is...
https://docs.circuitpython.org/en/latest/shared-bindings/board/index.html @lone axle the "board" module page is empty in the latest docs. not sure why.
it had correct content in 9.0.x https://docs.circuitpython.org/en/9.0.x/shared-bindings/board/index.html so this is a problem introduced in 9.1.0.
lone axle
strange, I'll try to look into that.
onyx hinge
OK, I can reproduce it locally, doing a bisect
commit 798e38e4470cbbc9316f40a9b161a37d9c8c52d4
Author: foamyguy <foamyguy@gmail.com>
Date: Sun Jun 23 11:14:35 2024 -0500
enforce min py3.9 for setboard. create init.py in board-stubs
Makefile | 1 +
tools/board_stubs/circuitpython_setboard/__init__.py | 4 ++++
2 files changed, 5 insertions(+)
I suspect the empty board/__init__.py created by touch is being used in lieu of the stuff in shared-bindings
onyx hinge
@lone axle thanks for bringing the problem up, I think I may have a solution ^
I don't think we have to cut an urgent release to fix this, because the intersphinx is just using the "latest" version, I think
onyx hinge
it's nice to be helpful when the rest of my day has consisted of little progress on my own tasks π happy I could help
kind river
My goal is a standard socket wrapper for the openthread TCP/UDP APIs (https://openthread.io/reference/group/api-tcp and https://openthread.io/reference/group/api-tcp)
slender iron
My goal is a standard socket wrapper for the openthread TCP/UDP APIs (https://op...
Great! Start with UDP. That's the land I'm in now. (Web workflow will need TCP though.)
kind river
That's the plan. UDP is generally preferred for openthread applications and should be easier to implement
kind river
Where should I inlude the openthread shared-bindings files?
slender iron
what should the module be called?
kind river
I'm just calling it openthread
It might make sense to have a generic thread module with an openthread implementation but I haven't seen a thread MCU that doesn't support openthread
slender iron
openthread is good. thread will conflict with the threading execution model
kind river
openthread is good. `thread` will conflict with the threading execution model
Yeah, many revised google queries have taught me that
especially before matter was a thing
kind river
should it live in shared bindings or shared modules? it could run on any port with a UART and an external radio but will have port specific init for ones with an internal radio
slender iron
the python to c layer is always in shared-bindings
the external uart impl could go in shared-module
willow totem
anyone aware of why the circuitpython-build-bundles CI job step might be throwing a wobbly:
https://github.com/adafruit/Adafruit_CircuitPython_AdafruitIO/actions/runs/9958291807/job/27513367239#step:2:767
spare jacinth
check out latest changes to the code
probably introduced something that is not supported on CP8
you can install the mpycross locally and test too
willow totem
I don't believe it's at the mpy-cross bit, it's only failing there due to failing the bundle script I think, or whatevers in the /tmp/randomName with the syntax error
spare jacinth
RuntimeError: ('mpy-cross failed on', '/home/runner/work/Adafruit_CircuitPython_AdafruitIO/Adafruit_CircuitPython_AdafruitIO/adafruit_io/adafruit_io.py')
and if you look slightly on top: Checking S3 for https://adafruit-circuit-python.s3.amazonaws.com/bin/mpy-cross/linux-amd64/mpy-cross-linux-amd64-8.2.0.static
willow totem
my mistake, it's also like 655 of my file not the tmp file
doesn't like this syntax self._post(path, {**metadata, "feeds": feeds_and_data})
spare jacinth
move the "feeds" before unpacking the dict, might just work
otherwise. runnin mpy locally will help identify/fix issues
(did for me, at least)
willow totem
nah tbf it likes the other order fine
onyx hinge
Traceback (most recent call last):
File "<stdin>", line 1
SyntaxError: invalid syntax
```this is not accepted by circuitpython, though it's valid in python 3.11.(and possibly earlier, 3.11 is the version I have handy)
{'feeds': 'whatever'}
```this syntax should workis the former invalid too if it has a leading normal keyvalue pair, in circuitpython?
spare jacinth
you could also some_dict | {"feeds": "value"} (at least on CPython)
onyx hinge
** unpacking within {} dict literals is not supported. I don't think it matters if there's a key-value pair before or after it.
onyx hinge
yes vertical bar syntax seems to work in circuitpython ```>>> {'a': 1} | {'b': 2}
{'a': 1, 'b': 2}
thanks folks, appreciate the schooling
onyx hinge
there's no way to know all these things in advance, but it's good to be aware there are Python language constructs not supported in CircuitPython, and then check by running on a device if it's convenient
spare jacinth
FYI, if im not mistaken, | (aka dict.__or__) is just an alias for dict.copy + dict.update
as in... could have used those manually if __or__ didnt work
willow totem
I assume the dict constructor version is doing similar under the hood. Mainly I didn't like the idea of an unnecessary for loop to update all the present keys. Mental note always try on device even if a PITA
spare jacinth
if you dont have to leave metadata untouched, could .update on it
and use it later
spare jacinth
I assume the dict constructor version is doing similar under the hood. Mainly I ...
yeah, your code above will unpack existing dict, so you end up with {"key": "value", "key2": "value2"} which is (almost?) equivalent to dict(key="value", key2="value2")
willow totem
I dump an array of dicts into one new field, but yeah similar. I like the metadata.update rather than a copy+update, a good suggestion
spare jacinth
beware you can't replace your code with metadata.update
it returns None, not the changed dict
because it edits inplace
thus my "only if you can afford to edit metadata" before
willow totem
metadata.update({"feeds": feeds_and_data})
self._post(path, metadata)
spare jacinth
does this look worthy of a PR to the bundle? https://github.com/elpekenin/circuitpython_backports
if so, ideas for anything else that might be a good and not too hard to implement addition?
onyx hinge
@spare jacinth Out of all those things, @contextmanager is the one I miss sometimes.
onyx hinge
D (31470) lwip: pbuf_alloced_custom(length=150)
D (31475) lwip: pbuf_free(0x3fff3834)
D (31479) lwip: pbuf_free: deallocating 0x3fff3834
D (31484) lwip: dhcp6_information_request: INFOREQUESTING -> -4
D (31490) lwip: DHCPv6 state: 1 -> 2 (dhcp6_information_request)
D (31496) lwip: dhcp6_information_request(): set request timeout 2000 msecs
LWIP_DEBUGF(DHCP6_DEBUG | LWIP_DBG_TRACE, ("dhcp6_information_request: making request\n"));
β¦ err = udp_sendto_if(dhcp6_pcb, p_out, &dhcp6_All_DHCP6_Relay_Agents_and_Servers, DHCP6_SERVER_PORT, netif);
pbuf_free(p_out);
LWIP_DEBUGF(DHCP6_DEBUG | LWIP_DBG_TRACE | LWIP_DBG_STATE, ("dhcp6_information_request: INFOREQUESTING -> %d\n", (int)err));
```It appears `udp_sendto_if` fails with `ERR_RTE = -4,` but of course there is not yet any ipv6 route when you're doing dhcp6!weird!
onyx hinge
('FE80::7EDF:A1FF:FE00:518C', 'FD5F:3F5C:FE50:0:7EDF:A1FF:FE00:518C', (nil), (nil), (nil))
```oh hey we're getting somewhereπ
onyx hinge
I can now attempt to connect to a v6 service, but the destination address is incorrect in the transimtted packet.```>>> s = socket.socket(socket.AF_INET6, socket.SOCK_STREAM)
s.connect(("fd5f:3f5c:fe50:0:6d9:f5ff:fe1f:ce10", 22))
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
OSError: [Errno 119] EINPROGRESS
```D (276442) lwip: IPv6 header:
D (276444) lwip: +-------------------------------+
D (276449) lwip: | 6 | 0 | 0 | (ver, class, flow)
D (276456) lwip: +-------------------------------+
D (276461) lwip: | 24 | 6 | 64 | (plen, nexth, hopl)
D (276468) lwip: +-------------------------------+
D (276473) lwip: | fd5f | 3f5c | fe50 | 0 | (src)
D (276478) lwip: | 7edf | a1ff | fe00 | 518c |
D (276483) lwip: +-------------------------------+
D (276488) lwip: | 3001 | 600 | 5da | 880 | (dest)
D (276494) lwip: | 26 | fe3f | 7085 | 23f |
D (276499) lwip: +-------------------------------+
onyx hinge
*** len[0]=28
*** len=28 family=10 port=5632
>>> s.recv_into(buf)
40
>>> bytes(buf)
b'SSH-2.0-OpenSSH_9.2p1 Debian-2+deb12u3\r\n\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00'
```OK I actually connected over v6 π really this timebut DNS messes up once I connect to v6... ```>>> wifi.radio.ipv4_dns
10.0.2.76
wifi.radio.ipv6_addresses
wifi.radio.start_dhcp()
wifi.radio.ipv4_dns
253.95.63.92
>>> s.connect(("fd5f:3f5c:fe50::1", 80))
>>> s.send("GET / HTTP/1.0\r\n\r\n")
18
>>> b = bytearray(100)
>>> s.recv_into(b)
100
>>> b
bytearray(b'HTTP/1.0 200 OK\r\nConnection: close\r\nETag: "31e-346-65e90ba6"\r\nLast-Modified: Thu, 07 Mar 2024 00:34:')
>>> s.close()
```you can manually make a request on it, but I think ipv6-address URLS like `http://[fd5f:3f5c:fe50::1]/` are probably not understood by adafruit_requests yet, and it needs to know whether to select `AF_INET` or `AF_INET6`.
onyx hinge
@slender iron ipv6 udp socket works, except the host name is not returned for recvfrom_into. raw ntp test: ```>>> ntp_addr = ("fd5f:3f5c:fe50::20e", 123)
PACKET_SIZE = 48
buf = bytearray(PACKET_SIZE)
with socket.socket(socket.AF_INET6, socket.SOCK_DGRAM) as s:
... s.settimeout(1)
... buf[0] = 0b0010_0011
... s.sendto(buf, ntp_addr)
... print(s.recvfrom_into(buf))
... print(buf)
...
48
(48, ('0.0.0.0', 123))
bytearray(b'$\x01\x03\xeb\x00\x00\x00\x00\x00\x00\x00GGPS\x00\xeaA0h\x07s;\xc0\x00\x00\x00\x00\x00\x00\x00\x00\xeaA0n\xeb4\x82-\xeaA0n\xebAU\xb1')
@onyx hinge nice! I think I will need the sender address too
I need to get back to sending packets today. TLV encoding tests are all green atm.
onyx hinge
very cool
manic glacierBOT
kind river
I'm getting a "CIRCUITPY_OPENTHREAD" is not defined error. Where should that be defined?
I added CIRCUITPY_OPENTHREAD = 1 to the Espressif port config which seems to be where the others are defined
spare jacinth
I'm getting a ```"CIRCUITPY_OPENTHREAD" is not defined``` error. Where should th...
probably, from a makefile
makefile != c compiler
so you are likely lacking something similar to: ```mk
C_FLAGS += -DCIRCUITPY_OPENTHREAD=$(CIRCUITPY_OPENTHREAD)
should be easy to find an equivalent configuration and write that correctly, rather than my assumption/guess
kind river
possibly, the only define I can find for wifi is in the port config
spare jacinth
not sure what you mean by that
kind river
spare jacinth
but sounds like you were just missing the "glue" from makefile to C (or incorrectly assuming they were implicitly/inherently glued by default - they are not)
kind river
I'm trying to figure out where it is
spare jacinth
search for $(some_other_config)
kind river
found it
spare jacinth
so you are likely lacking something similar to: ```mk
C_FLAGS += -DCIRCUITPY_OPE...
but again.. should be quite similar to this
-DVALUE=ERROR being passed to the compiler is equivalent to having #define VALUE ERROR on each and every compilation unit
ornate breach
Whatβs the likelihood that getting an IMX RT1176 to work would be fairly straightforward going off of the current NXP RT implementation for circuitpython?
tulip sleet
Whatβs the likelihood that getting an IMX RT1176 to work would be fairly straigh...
I think it's mostly a more featureful version of the other RT11xx, so it might be relatively straightforward.
making the second core available in some way is more work
manic glacierBOT
slender iron
Whatβs the likelihood that getting an IMX RT1176 to work would be fairly straigh...
I tried ages ago but it was unreliable.
@ornate breach https://github.com/tannewt/circuitpython/commits/imxrt11xx/
kind river
I'm thinking of having an openthread.PendingDataset singleton that wraps a dataset and has an openthread.PendingDataset.SetSctive() function that sets the wrapped dataset as active
I could allow creation of multiple dataset instances but that seem overly complex
ornate breach
<@728676724522352733> <https://github.com/tannewt/circuitpython/commits/imxrt11x...
what was the biggest issue with the port? Is it something that could be managed?
Or do you have a list of things that need to be resolved to make it reliable?
shadow vigil
dumb question about class inheritance in CircuitPython -- i've got a 3-deep inheritance that the super-classes own attributes, i get an "AttributeError: 'super' object has no attribute -- is this kind of an expected thing due to the limited memory constraints?
(i don't like posting bugs/issues for things that are "yeah, we don't do that" type items)
onyx hinge
@shadow vigil there are some limitations of super() compared to standard python. the one I remember is super() & properties do not work as standard python. and lots of built in classes do not properly support subclassing. e.g., this from micropython docs. we inherit most of the language differences/limitations from micropython, as we don't do a whole lot of work on the core language implementation. https://docs.micropython.org/en/latest/genrst/core_language.html#calling-super-getter-property-in-subclass-will-return-a-property-object-not-the-value
shadow vigil
apparently time for plan b -- once i figure out what plan b is π
sleek relic
apparently time for plan b -- once i figure out what plan b _is_ π
Property annotations are a 'convenience'. Try without the annotations and use something like .get_prop_name() [or just .prop_name()] instead .prop_name.
and .set_prop_name(value)
sleek relic
For closer to the functionality in the referenced page try this (raw example, not tested in any way):
class A:
def __init__():
self._p = None
def _set_p(self, value):
self._p = value
def _get_p(self):
return self._p
class AA(A):
@property
def p(self):
return self._get_p()
The non-property methods should still inherit, so that structure should work. No super reference needed unless functionality needs to be overridden.
manic glacierBOT
shadow vigil
Property annotations are a 'convenience'. Try without the annotations and use so...
thanks, but that's not actually the point -- it's "depth of inheritance": a parent with a child class is fine, a "grand-child" doesn't play nice -- otoh, i'm "only" using single inheritance, so...
slender iron
what was the biggest issue with the port? Is it something that could be managed?
Its hard to remember. You'll have to try it and see
ornate breach
Okay, itβll be a bit. I actually have to design the hardware first lol.. rt1176 feather π¬
slender iron
I started a 1060 feather. that's what I'd work on
lilac scaffold
hello! Looking for help: I have a custom PCB with the STM32H743 on it. Having a hard time getting it to boot into CircuitPy drive.
I can pull BOOT0 High and get it into USB/ DFU but touching the NRST pin to GND does not bring up the drive. I'm on Mac.
Default BOOT0 is low.
slender iron
hello! Looking for help: I have a custom PCB with the STM32H743 on it. Having a ...
π We don't have any adafruit boards with an H7 so it may be broken unfortunately. I'd suggest connecting a debugger and getting a backtrace.
lilac scaffold
I've read that there might be an issue with the bootloader (I got mine from Nucleo) - so I uploaded v7.0. The latest version is 9.0
slender iron
which bootloader?
lilac scaffold
π We don't have any adafruit boards with an H7 so it may be broken unfortunatel...
π this bootloader https://circuitpython.org/board/nucleo_h743zi_2/ but then I went here https://adafruit-circuit-python.s3.amazonaws.com/index.html?prefix=bin/nucleo_h743zi_2/en_US/ and got adafruit-circuitpython-nucleo_h743zi_2-en_US-7.0.0.bin
I have a STMLink, will that work as a debugger?
slender iron
ah, that isn't a bootloader. that is the circuitpython binary
yes STMLink will work as a debugger
lilac scaffold
oh! Am I uploading the wrong thing to the board? I'm trying to see a CircuitPy drive
slender iron
nope, that seems like the right thing
it just isn't a bootloader (a bootloader is separate software that facilitates loading other software)
tinyuf2 is an example of a bootloader
I think your best bet would be to build circuitpython yourself and then use a debugger
lilac scaffold
so basically build it on mac, upload using STLink, then debug w STLink?
slender iron
yup. once you build it, then you can add blinks into the early code to see if it gets that far
you'll also have debug symbols
this channel is the best place to get help π
lilac scaffold
I came immediately to the right place then. Yay!
onyx hinge
>>> import wifi
>>> wifi.radio.ping("1.1.1.1")
0.021
>>> wifi.radio.start_dhcp()
>>> wifi.radio.addresses
('FE80::7EDF:A1FF:FE00:518C', 'FD5F:3F5C:FE50:0:7EDF:A1FF:FE00:518C', '10.0.3.96')
>>> wifi.radio.dns
'FD5F:3F5C:FE50::1'
>>> wifi.radio.ping(_)
0.014
@slender iron recvfrom & accept, which return different data when v6 packets are in use, kinda break the C API... cpython 3.11: ```>>> s.recvfrom(len(packet))
(b'$\x01\x03\xeb\x00\x00\x00\x00\x00\x00\x00DGPS\x00\xeaB\x8c\xf0\x07y\x93\x15\x00\x00\x00\x00\x00\x00\x00\x00\xeaB\x8c\xf3\xc7"\xf7l\xeaB\x8c\xf3\xc7-\x86\xae', ('fd5f:3f5c:fe50:0:3ca7:8912:9d4:8360', 123, 0, 0))
vs a 2-tuple. ```>>> s.recvfrom(len(packet))
(b'$\x01\x03\xeb\x00\x00\x00\x00\x00\x00\x00DGPS\x00\xeaB\x8d8\x07u\xda\xee\x00\x00\x00\x00\x00\x00\x00\x00\xeaB\x8d;\xe1\xe6\xad]\xeaB\x8d;\xe1\xf5\x1d+', ('10.0.2.74', 123))
I can trample all over it, just letting you know.
slender iron
np, do what must be done π
shadow vigil
<@1094335828450353242> there are some limitations of super() compared to standar...
GAH! I figured it out -- it was a very weird sort of circular reference that was pretty easy to fix by over-riding methods instead of super() calls
misty garnet
the more I've used inheritance, the less I've liked it
shadow vigil
it works in languages built for it π
sleek relic
thanks, but that's not actually the point -- it's "depth of inheritance": a pare...
Nothing in the reference document said there were limitations for multiple levels of inheritance. It only showed that @property does not play nice with super(). Can you post sample code with grandchild class that does not do what you expect?
shadow vigil
Nothing in the reference document said there were limitations for multiple level...
silly names and all π -- but i did realize that i was really not doing this very pythonic, and it was a very poor translation from Kotlin (yep AI strikes again)
sleek relic
Using CPython 3.10.14 and CircuitPython 9.5
Output from your sample code in CPython
Multi-inheritance
static one
middle {'state': 'yah, nope', 'static': 'one'}
upper: {'state': 'yah, nope', 'static': 'one', 'okay': 'cool'}
{'state': 'yah, nope', 'static': 'one', 'okay': 'cool'}
CircuitPython fails with AttributeError.
My Modified code, avoiding use of inherited @property, methods gives the identical result with CPython, and "equivalent" with CircuitPython. Equivalent not identical, because the dictionary elements are being reported in a different order, although the values match. Key reference order is not guaranteed in python for regular dict data type.
Multi-inheritance
static one
middle {'state': 'yah, nope', 'static': 'one'}
upper: {'static': 'one', 'state': 'yah, nope', 'okay': 'cool'}
{'static': 'one', 'state': 'yah, nope', 'okay': 'cool'}
My version of the code. The statical properties of MostbasestClass and MiddleThingie are not needed for this sample, but I left them in, in case instances of those classes want to use them.
class MostbasestClass:
def __init__(self, static_value:str, other_value:str):
self._static_value = static_value
self._state_value = other_value
@property
def state(self):
return self._state_value
@state.setter
def state(self,value:str):
self._state_value = value
def _statical(self)->dict:
print(f"static {self._static_value}")
return {
"state":self._state_value,
"static":self._static_value
}
@property
def statical(self)->dict:
return self._statical()
class MiddleThingie(MostbasestClass):
def __init__(self, static_value:str, other_value:str):
super().__init__(static_value, other_value)
self._yadda = "cool"
def _statical(self)->dict:
v = super()._statical()
print(f"middle {v}")
v["okay"] = self._yadda
return v
@property
def statical(self)->dict:
return self._statical()
class HereItIs(MiddleThingie,MostbasestClass):
def __init__(self, static_value:str, other_value:str):
super().__init__(static_value, other_value)
@property
def statical(self)->dict:
v = self._statical()
print(f"upper: {v}")
return v
print("Multi-inheritance")
h = HereItIs("one","two")
h.state = "yah, nope"
print(h.statical)
HI - I'm building circuitpython and getting an error at make -C mpy-cross, have tried many things.
spare jacinth
and error being....?
it works for me running the exact same command from the circuitpython folder... you might need some dependency (perhaps from make fetch-all, perhaps some python lib or somthing, idk)
manic glacierBOT
lilac scaffold
- Verbosity options: any combination of "steps commands rules", as `make V=...` or env var BUILD_VERBOSE
Cannot determine version.
CircuitPython must be built from a git clone with tags.
If you cloned from a fork, fetch the tags from adafruit/circuitpython as follows:
make fetch-tags
make: *** [build/genhdr/mpversion.h] Error 1```did you do make fetch-tags at the top level of the repo?
lilac scaffold
by top level, do you mean above the circuitpython folder that I cloned into?
tulip sleet
in the circuitpython folder. Looks like you are there now
lilac scaffold
yes I did
lilac scaffold
git fetch --tags --recurse-submodules=no --shallow-since="2023-02-01" https://github.com/adafruit/circuitpython HEAD
remote: Total 0 (delta 0), reused 0 (delta 0), pack-reused 0
From https://github.com/adafruit/circuitpython
* branch HEAD -> FETCH_HEAD```yes I still get the error w mpy
I'm following along here https://learn.adafruit.com/building-circuitpython/build-circuitpython
tulip sleet
how did you do the clone?
lilac scaffold
git clone --recurse-submodules https://github.com/adafruit/circuitpython.git
I've re-cloned several times and still get the mpy error. I'm on an M2 mac, but that shouldn't matter right?
tulip sleet
`git clone --recurse-submodules https://github.com/adafruit/circuitpython.git`
you don't need to do --recurse-submodules. We have some make targets for that, that only fetch what you need
what version of macOS are you using?
what does git --version say?
lilac scaffold
I'm on Ventura, git version 2.45.2
tulip sleet
that should be ok
are you using a case-sensitive filesystem, btw?
not sure that's the problem here, but if not, it may be a problem later
tulip sleet
have you read https://learn.adafruit.com/building-circuitpython/macos and the rest of that guide?
lilac scaffold
no, I went straight to building it!
tulip sleet
we have pre-built mpy-cross, by the way, here:
https://adafruit-circuit-python.s3.amazonaws.com/index.html?prefix=bin/mpy-cross/
please do read the guide, there are many prep steps
lilac scaffold
thanks very much!
onyx hinge
web workflow via ipv6
lilac scaffold
have you read <https://learn.adafruit.com/building-circuitpython/macos> and the ...
I did read that guide, several hours ago. Anyway, I'm starting over. I have the partition and gmake 4.4.1
tulip sleet
i did a fresh clone to check and the right tags are there: git tag will show them. So the problem is the mpy-cross compile.
tulip sleet
I did read that guide, several hours ago. Anyway, I'm starting over. I have the ...
you may need to say gmake -C mpy-cross instead of make ...
as mentioned in the guide:
If you install the brew version of make, it will be known as gmake.
lilac scaffold
YES. I forgot to rename it. OK, will try that! thank you thank you!
that worked hurrah!
tulip sleet
I would say don't rename, just use gmake as the command name. It will propagate down.
anyway, you are on track!
lilac scaffold
ok! well, I'm stuck again. I'm trying to build for a chip that Adafruit doesn't use. Its the STM32H743, used on Nucleo
I see the chip in /boards.
make: *** No targets specified and no makefile found. Stop.```did you cd ports/stm ?
lilac scaffold
yea
tulip sleet
you have to pick a diretory name. the .ld file is not a board name
lilac scaffold
so I cd ports/stm/boards and tried nucleo_h743zi_2 also
lilac scaffold
make: *** No targets specified and no makefile found. Stop.
sophikravitz@MacBook-Pro-7 boards % make BOARD=openmv_h7
make: *** No targets specified and no makefile found. Stop.```do make fetch-port-submodules first, in ports/stm
do gmake not make
rename it back
lilac scaffold
good looking out, appreciate it
tulip sleet
move out of the boards directory
lilac scaffold
yup
tulip sleet
up one, to ports/stm
tulip sleet
are you doing your own build to adjust the build?
since there's a build here already: https://circuitpython.org/board/nucleo_h743zi_2/
there is no UF2 bootloader for these boards
you'll have to use the .bin and the ROM bootloader on the board
lilac scaffold
I have a custom board and the bin from Nucleo isn't showing me CircuitPy drive.
I'm able to connect using STLink and also via USB but the board just wont boot into CircuitPy
tulip sleet
we haven't tested these builds for a while, so you could try older ones here: https://adafruit-circuit-python.s3.amazonaws.com/index.html?prefix=bin/nucleo_h743zi_2/
lilac scaffold
I tried 7.0.0 earlier from this link, I think the latest is 9.0.0
Maybe its my hardware then
Altho, everything else seems to be ok
tulip sleet
do you have a nucleo also to test with?
lilac scaffold
no, but I'll get one. I have an open MV here, but it uses the F version of the H chip
slender iron
you can add DEBUG=1 to do a debug build with symbols too
austere acorn
no, but I'll get one. I have an open MV here, but it uses the F version of the H...
unfortunately, the nucleoh743 is obsolete, can't buy it anymore
lilac scaffold
you can add `DEBUG=1` to do a debug build with symbols too
where do I add that? just after BOARD= ?
slender iron
yup
austere acorn
i wanted to buy that one too, cos i suspect the builds don't work for the h7
lilac scaffold
just dropped the 5.1 version from Nucleo and it's not working
so maybe there's an issue with the STM32H743 build. Someone on the STM forum thought that there was
I have never used Circuit Python π€·
tulip sleet
I think that build was done by a volunteer. We don't sell a board with that chip, so it's not high priority.
austere acorn
The stm line of chips definitely needs more love in circuitpython π
lilac scaffold
I'm going to try the H7 on the open MV now
Do you know if it's ok to download it over USB?
tulip sleet
you need to use the dfu loader, I assume.
MicroPython has STM32 support, and it was originally for STM chips, so it may have better support. We won't feel bad if you go that direction π
austere acorn
I'm going to try the H7 on the open MV now
You have the OpenMV Cam H7 ? Could you let me know if the circuitpython image works on it?
That's the only other port using H7, and I'm trying to confirm a theory... but I don't have the board.
kind river
9 doesn't work on my Nucleo H743ZI
austere acorn
9 doesn't work on my Nucleo H743ZI
Does 8? Or other previous versions?
shadow vigil
My version of the code. The statical properties of MostbasestClass and MiddleThi...
eh, i found a cleaner way of doing what i mostly wanted that wasn't so convoluted -- i realised that the code that was translated by one of the AI generators was messing with me
kind river
Does 8? Or other previous versions?
6 works
manic glacierBOT
lilac scaffold
You have the OpenMV Cam H7 ? Could you let me know if the circuitpython image w...
I'm not able to connect to the H7 so I don't think it works
slender iron
I'm not able to connect to the H7 so I don't think it works
Try CP 6 on your board
lilac scaffold
Try CP 6 on your board
The openMV variant that I have doesn't connect to dfu at all / is not recognized by computer - it's OpenMV3 R2 CN-ET-003 OV7725-M7
btw the STM32H743 doesn't connect on Micropython either.
Maybe it's my hardware, I forgot a pullup on DP and soldered it on by hand. If hw is the issue, I'm not sure why it connects on dfu tho.
slender iron
dfu over usb should indicate usb is ok
although maybe that runs as usb full speed (can it do high speed otherwise?)
lilac scaffold
I don't know. The only thing I've been able to verify without code is the power and dfu.
lilac scaffold
I have a stlink π can you point me to docu on this?
slender iron
and gives you another way to load code besides dfu
lilac scaffold
yes, but only code in C/C++ right?
slender iron
ya, some native code
it should help you get circuitpython going though
https://learn.adafruit.com/debugging-the-samd21-with-gdb is my process with jlink
I've not used stlink myself
lilac scaffold
cool, I'll give it a shot.
Still, if micropython isn't working either...
slender iron
I'd expect micropython st support to be working
lilac scaffold
the issue is the hw- I thought that CP communicates over DP and DM. π
onyx hinge
@slender iron can you briefly review why IP address objects were useful e.g., with ping? in veesix I did NOT add an IPV6Address object, and I made ping able to take an IPv4/IPv6 string (or actually anything resolvable with getaddrinfo).. I want to make sure this is OK.
plucky tulip
Hi, is there maybe someone who owns Adafruit MEMENTO and would be able to help me test some code?
I do not have one myself, and I would like to make sure that the example I want to create for adafruit_httpserver works before creating a PR.
If anyone would be able to help, please DM me. π
manic glacierBOT
willow totem
Hi, is there maybe **someone who owns Adafruit MEMENTO** and would be able to he...
Maybe can help you @plucky tulip probably tomorrow night or over weekend. Hopefully someone helps before then
What time zone you on?
plucky tulip
Maybe can help you <@699200371909853216> probably tomorrow night or over weekend...
Thank you. It is not urgent, also it probably won't take much time, I hope. I am in Europe/Warsaw CEST.
Over weekend would be great.
willow totem
foamyguy / others: Apologies I've probably asked this before, but is there anything I should be aware of when doing release notes or even releases in general for circuitpython libraries? Just going to push out some changes for Adafruit IO circuitpython library. I'd probably just use auto generate notes and then tweak unless told otherwise.
plucky tulip
Maybe can help you <@699200371909853216> probably tomorrow night or over weekend...
What time would work for you? I will try to adapt
willow totem
To be honest no plans yet, most likely I'll be online around when the deep dive stream happens and Tim's Saturday stream at about 4 I'm often around
plucky tulip
Great, I will prepare everything before then. Thanks
tulip sleet
foamyguy / others: Apologies I've probably asked this before, but is there anyth...
For libraries, we almost always used the autogenerated notes. The release name should include the version number and a brief summary, e.g. `3.2.1 - Allow busy-wait loop in send to be interrupted"
manic glacierBOT
willow totem
For libraries, we almost always used the autogenerated notes. The release name s...
Cool, thanks Dan. Sounds similar to my own thoughts. Only tricky bit is squeezing 3 features into one release title. I usually pick one, or cram two occasionally ποΈ or go more generic.
manic glacierBOT
slender iron
<@252717193496756235> can you briefly review why IP address objects were useful ...
I think I was mimicking: https://docs.python.org/3/library/ipaddress.html
onyx hinge
OK so an IPv6Address object might be useful as a way for the Python program to interact with the byte-wise data in an ipv6 address...
onyx hinge
@slender iron should I have start_dhcp_client_v6 and start_dhcp_client_v4? with the existing start_dhcp_client .. starting both? and both starting by default when bringing up the interface?
slender iron
maybe start_dhcp_client with two kwargs that default to true?
slender iron
sure
onyx hinge
thanks!
digital shoreBOT
adafruit
kind river
I am having issues with ```#
OpenThread
CONFIG_OPENTHREAD_ENABLED=y
end of OpenThread```
not being copied from sdkconfig.defaults to the generated esp-idf config. I"ve tried running make clean and deleting the build directory
kind river
make menuconfig has no openthread options despite the openthread component existing in the esp-idf source folder
willow totem
Great, I will prepare everything before
onyx hinge
I wonder if https://github.com/adafruit/circuitpython/pull/9441 has anything to do with @lone axle 's BLE problems trying to do circup file transfers
tulip sleet
I wonder if https://github.com/adafruit/circuitpython/pull/9441 has anything to ...
this was a specific problem with nordic. Were those problems with espressif or nordic?
onyx hinge
I guess I don't know. I just recall that some simple operations would work but bigger operations would hang for no clear reason.
I don't know which MCU tim was testing with, either
tulip sleet
i have other known problems to debug now too π , so I can look at those
kind river
make menuconfig has no openthread options despite the openthread component exist...
Any ideas on why esp-idf seems to be ignoring the openthread component?
tulip sleet
Any ideas on why esp-idf seems to be ignoring the openthread component?
I tried finding it too, and could not. There is an openthread/Kconfig. I tried "A" ahd "/" in menuconfig, and it couldn't find anything. I looked for a list of components somewhere and couldn't find that either. I did some websearching about this but that was also unusccessful.
kind river
I tried finding it too, and could not. There is an openthread/Kconfig. I tried "...
Thanks for trying, at least I know it's not just me now. That's pretty much what I tried too. I also tried adding ESP_IDF_COMPONENTS_LINK += openthread to ports/espressif/Makefile with no noticeable change.
tulip sleet
@stuck elbow hi, a user wrote to us and noted that http://circuitpython-stage.readthedocs.io/ is broken right now. The builds are failing for some reason. The link from https://learn.adafruit.com/circuitpython-stage-game-library/documentation is how they found that out.
stuck elbow
<@351136009922019331> hi, a user wrote to us and noted that http://circuitpython...
I noticed it a week ago too, but I've been down with covid since. I will look into it as soon as I can. Thank you for the notice.
stuck elbow
fixed
tulip sleet
I noticed it a week ago too, but I've been down with covid since. I will look in...
Thank you! I hope you are feeling better.
lone axle
I wonder if https://github.com/adafruit/circuitpython/pull/9441 has anything to ...
π I will try this out on a newer version from after this PR is merged. Thank you!
I'll try to distill it into a smaller reproducer if I can once I get back to it as well. If I manage that I'll submit an issue with it. Though some part of me fears that I'll strip out the layers of Click and the circup "Backend" architecture and then not have the same problem.
onyx hinge
"removed definition of board_display_obj" is a pygamer-specific change after 9.1.0-beta.3
but that removal looks so sensible
lone axle
I noticed an issue other the weekend for a DisplayIO Driver Library (SSD1322) that wasn't updated for the newer displayio API changes where displayio.Display became busdisplay.BusDisplay
I checked the rest of the bundle and found there are 4 such drivers in need of updates still and a small handful of helper libraries that need docstrings and/or type annotation updated but should function AFAIK.
Looking into that this morning I've found we have 8.x / 9.x compatibility code in a number of the other display driver libraries i.e. https://github.com/adafruit/Adafruit_CircuitPython_SSD1325/blob/main/adafruit_ssd1325.py#L33
Is it okay to go ahead and remove that compatibility fallback and just update to the newer code? If so I can submit PRs today for that change along with the fix for the ones using the old version only.
onyx hinge
I think 8.x is still "in the bundle" so probably not yet.
@lone axle thanks for the sleuthing about the pyportal/pygamer display problem. do you want to continue by "bisect"ing or is that as far as you go?
lone axle
<@382939733107408897> thanks for the sleuthing about the pyportal/pygamer displa...
I'd be happy to bisect it further between beta.3 and beta.4 to try to find the one where it changes.
onyx hinge
>>> wifi.radio.ping("ipv6.google.com")
0.048
>>> wifi.radio.addresses
('FE80::7EDF:A1FF:FE00:518C', '2001:470:3996:0:7EDF:A1FF:FE00:518C', '2001:470:7C:285:7EDF:A1FF:FE00:518C', '10.0.3.96')
I forget who reminded me of the existence of ipv6 tunnels so that I could test this
but hug report to you
slender iron
@kind river did you add it to the CP CMakeLists file?
orchid basinBOT
slender iron
<@&356864093652516868> Just under an hour until our meeting. Excited to chat with you all. Notes doc is here: https://docs.google.com/document/d/1cvCxFaJT5hIQ4UmzZm6f_betgOKNhtXJrsA4GYdVeQ4/edit?usp=sharing
onyx hinge
huh v6 support might only cost around 32kB, possibly actually feasible to add to pico w if it's that small.
tulip sleet
@slender iron https://github.com/adafruit/circuitpython/commit/109bcafd072bd94a9eef4e4d06657efa4a112b0e#diff-d695723d1b6e886103267211b0f1867a65f2e83cf0e8fb1ac639c54828550859 -- I don't see a PR with this change
**to all folks making PRs: **I just created branch 9.1.x on adafruit/circuitpython. If you are submitting a bugfix PR, do it against 9.1.x. New features can go against main.
slender iron
<@252717193496756235> https://github.com/adafruit/circuitpython/commit/109bcafd0...
yup, I must have forgotten
tulip sleet
yup, I must have forgotten
we can do this against 9.1.x now. I can make a new PR. I will check the other inky frame too. Don't bother to make a PR.
slender iron
π thanks!
onyx hinge
fan noise seems fine, did hear the sirens though
lone axle
onyx hinge
π
lone axle
random junco
I've been playing around with ViperIDE this weekend, it's pretty neat
lone axle
Pypi is optional
When you run cookie cutter it will ask you yes or no if you want to deploy to pypi
dense glacier
https://learn.adafruit.com/creating-and-sharing-a-circuitpython-library
I was looking to create a library but found someone else beat me to it
https://github.com/iorodeo/iorodeo-as7331
Thanks IORodeo for this. Worked great for UV sensor
lone axle
Thanks for hosting Scott. Have a great week everyone.
slender iron
np!
slender iron
Here is the notes document for next Mondayβs CircuitPython Weekly Meeting. It is at the normal time of 11am Pacific / 2pm US Eastern here on Discord. Add your hug reports and status updates to the document before the meeting. If you are unable to attend but would still like to contribute, feel free to add your notes and weβll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/1B3UNUMkUamVZ9-aSAbCf8csy48K5ER-eUjna9yU6Arc/edit?usp=sharing
slender iron
Here is the notes document for next Mondayβs CircuitPython Weekly Meeting. It is...
slender iron
@lone axle I think it is too early to switch to the split displayio layout in libraries. Blinka displayio tries to support it but isn't quite right
BusDisplay needs to be updated to import BusDisplay as Display
lone axle
<@382939733107408897> I think it is too early to switch to the split displayio l...
Most of the display drivers have code that is compatible with both BusDisplay and displayio.Display: https://github.com/adafruit/Adafruit_CircuitPython_SSD1325/blob/main/adafruit_ssd1325.py#L33
Do we want to have that same compatible code in the 4 that don't have it currently?
Or just leave it as-is for now, but I think if we don't make a change anywhere then those 4 display drivers are un-usable in some contexts, Blinka_DisplayIO I think, but still work in the core perhaps.
A user reporting not being able to use one of those display drivers here https://github.com/adafruit/Adafruit_CircuitPython_SSD1322/issues/21 which is what led me down this path. I'm happy to do whatever remediation is best at this time.
slender iron
I think fixing blink_displayio would be my preference. that will make it match CP 9
kind river
<@177230548236763136> did you add it to the CP CMakeLists file?
Thanks for the tip! The component is being included now but openthread/dataset_ftd.h still isn't found.
slender iron
did that fix menuconfig? to fix headers you will need to edit the Makefile where it has -isystem paths
(we don't use cmake for the CP side of the build)
@onyx hinge do you have avahi working but not .local lookup in your browser?
onyx hinge
@slender iron I don't really know what the deal is. Trying to use .local has never worked reliably for me, so I mostly don't even try anymore. /shrug
slender iron
avahi has a step to connect it to domain lookup on arch: https://wiki.archlinux.org/title/Avahi#Hostname_resolution I use resolved for domain look up so it does the mdns queries
slender iron
@onyx hinge I think this is waiting your follow up review: https://github.com/adafruit/circuitpython/pull/9427
onyx hinge
looking, thanks
slender iron
I'll review your ipv6 changes next
kind river
did that fix menuconfig? to fix headers you will need to edit the Makefile where...
Menuconfig is working but adding -isystem esp-idf/components/openthread/include doesn't seem to
kind river
shared-bindings/openthread/PendingDataset.h
slender iron
shared-bindings/openthread/PendingDataset.h
hrm. I'd expect adding the -isystem addition to work. maybe the path isn't exactly right? I use fd to find where a header is in the esp-idf tree
orchid basinBOT
onyx hinge
@slender iron maybe you can decide whether to delete or document, I created a PR #9446 to duel with #9427 π
slender iron
k, let me finish your review
orchid basinBOT
kind river
The actual openthread libs were in openthread/openthread/include. Now I'm getting error: 'mbedtls_ssl_set_hs_ecjpake_password' was not declared in this scope
slender iron
that may be behind a setting
lone axle
Does anyone know if it's possible or how to specify requirements for a cookiecutter template?
In one of our cookiecutter hooks we use pkg_resources for parsing semantic versions: https://github.com/adafruit/cookiecutter-adafruit-circuitpython/blob/251be2cfd205b427ae7807d08a40695ef3353b04/hooks/pre_gen_project.py#L6-L12 but in python 3.12 this is deprecated or removed and doesn't work (at least in my environment) so the cookiecutter fails after asking all of it's prompts.
I have found that https://pypi.org/project/packaging/ exists and seems to implement the same functionality so in my local branch I've changed the code to use that and it works if I manually pip install packaging but ideally we'd want cookiecutter to install that whenever the user runs cookiecutter gh:adafruit/cookiecutter-adafruit-circuitpython I tested it with my own branch and it fails if I don't manually install that package. I tried adding a requirements.txt thinking that might do it, but it doesn't seem to have made a difference.
Or as an alternative the code could be simplified to parse "manually" with string split and int() and then compared to the minimum version. As far as I understand it this code looks to just be enforcing a minimum version which does feel like a bit minimal to be requiring an external module for.
kind river
yep, menu config option for mbedtls. everything is compiling now
lone axle
Honestly I don't think that mechanism for enforcing the minimum version is actually effective. I installed cookiecutter 1.7.3 which is lower than 2.1 and when I try to run it I get a stacktrace explosion about a filename being too long and it never prints the error from our hook.
crimson ferry
To start up code.py with IPv4 only, are these the required steps (with PR#9436?```
import wifi
wifi.radio.stop_dhcp_client()
wifi.radio.start_dhcp_client(ipv4=True, ipv6=False)
I have to admit I did not test this. In principle, just the 2nd call is needed, because it stops whichever one(s) are not requested. ```c
void common_hal_wifi_radio_start_dhcp_client(wifi_radio_obj_t *self, bool ipv4, bool ipv6) {
if (ipv4) {
esp_netif_dhcpc_start(self->netif);
} else {
esp_netif_dhcpc_stop(self->netif);
}
#if LWIP_IPV6_DHCP6
if (ipv6) {
esp_netif_create_ip6_linklocal(self->netif);
dhcp6_enable_stateless(esp_netif_get_netif_impl(self->netif));
} else {
dhcp6_disable(esp_netif_get_netif_impl(self->netif));
}
#endif
}
π the increasingly inaccurately named `start_dhcp_client`https://github.com/adafruit/circuitpython/pull/9436/files#diff-45099aee9920055cc86eca6488ba14d43010cf831c6083a9da55680ef8dc5fecR696 uhmmm I shoulda finished writing the docs I guess
crimson ferry
do we know if the espressif implementation uses the RFC 4941 βprivacy extensions"?
if not, I would suggest not starting IPv6 by default
onyx hinge
I do not have a strong opinion about what the default should be. Now is the time to influence that choice. I do not see any signs via full text search of rfc 4941 support, though I also don't know in detail what that is (I know it has to do with not linking global IPv6 address to hardware address, but that's just a very general level of knowledge)
crimson ferry
I don't have an in-depth understanding of it either, but I am concerned (and suspect some others would be as well) if their microcontrollers are permanently and globally identifiable
onyx hinge
I don't have an in-depth understanding of it either, but I am concerned (and sus...
Please raise this on the PR even if it's just a quick comment!
kind river
Almost there, getting undefined references to the coex functions
midnight ember
The symbol font Fork Awesome that is used in some Adafruit guides and that I was introduced to via Adafruit devs is unfortunately showing a deprecated notice on their Github. https://github.com/ForkAwesome/Fork-Awesome I'm not sure how that might affect some Adafruit projects or if the project will be deleted someday. Might be worth a dev having a quick look into it to see if there are any implications for adafruit guides. I have been using the font for some of my recent projects and I don't even know what it means. As long as the font file is included in the bundle for any learn guides then it should continue to function normally, my guess is there simply will not be anymore future revisions to fork awesome.
orchid basinBOT
tulip sleet
https://blog.adafruit.com/2024/07/22/circuitpython-9-1-1-released/
A few fixes, most notably backlight fixed on SAMD51 display boards.
onyx hinge
thanks for making that release @tulip sleet -- good to have the display big fixed
tulip sleet
thanks for making that release <@329766224093249548> -- good to have the displa...
you're welcome - it was a bad thing to have broken boards
wraith crow
Well I didn't expect git to interpret that update as a review dismissal π€·
tulip sleet
Well I didn't expect git to interpret that update as a review dismissal π€·
any time you push something it will dismiss an approval (since there's a change)
manic glacierBOT
onyx hinge
@slender iron do you want ipv6 to default on or off? and did you want the "supports_ipvx" constants removed? I'm eager to update that PR so it can be reviewed again.
slender iron
<@252717193496756235> do you want ipv6 to default on or off? and did you want th...
default off is fine with me if there are security concerns. ya, I'd remove the constants because they don't match CPython
onyx hinge
default off is fine with me if there are security concerns. ya, I'd remove the c...
I can do that -- but did you catch that the constants are on the wifi module not the socketpool module/object?
(the circuitpython enabling define says SOCKETPOOL but it affects stuff across wifi & socketpool both)
slender iron
ya. I think it'd be better to test in a way that matches CPython
onyx hinge
OK
onyx hinge
thank you for the merge @slender iron ! I look forward to seeing what you can get done with Matter over the next weeks while I'm away.
tulip sleet
@slender iron @onyx hinge shall I make a 9.2.0-beta.0 yet?
slender iron
how is the MP merge coming? I'm not sure 9.2 has much for folks to test in it
tulip sleet
good point, I should probably wait for that at least. Not even ready to try to compile yet.
I want to fix some more bugs too
slender iron
I'm not ready to run the matter stuff on CP anyway
kind river
the openthread binding is compiling but crashing due to memory access. need to brush up on my C
manic glacierBOT
empty salmon
For updates to an existing board, which branch should we fork from? main ? or 9.1.x ?
slender iron
9.1 if you want it released as stable soon
empty salmon
Thanks. I will fork branch from the 9.1.x for this minor change.
tulip sleet
Thanks. I will ~~fork~~ branch from the 9.1.x for this minor change.
and then set 9.1.x as the base in the PR if it does not do that on your PR push
@slender iron were you thinking of reviewinghttps://github.com/adafruit/circuitpython/pull/9453 or should I just merge it?
slender iron
I don't need to review it
analog pivot
Not sure if this belongs here, but I'm having trouble being able to copy into and delete files from my CIRCUITPY folder on my Macropad. How do I fix it? I am running MacOS SOnoma 14.5
stuck elbow
Not sure if this belongs here, but I'm having trouble being able to copy into an...
I think that should go to #help-with-circuitpython, but the short answer is that it's up to Apple to fix it
manic glacierBOT
kind river
I can create an instance of { mp_obj_base_t base; otInstance *instance; otOperationalDataset *dataset; } openthread_pending_dataset_obj_t; but when I try to access it, I get Hard fault: memory access or instruction error.
(on the C side)
tulip sleet
I can create an instance of ```{
mp_obj_base_t base;
otInstance *instanc...
how do you create the instance?
kind river
how do you create the instance?
openthread_pending_dataset_obj_t openthread_pending_dataset_obj;
tulip sleet
```openthread_pending_dataset_obj_t openthread_pending_dataset_obj;```
and how are you accessing it? Are you copying some idiom from somewhere else?
you could point me to your repo
tulip sleet
got it, so where are you getting the hard fault?
kind river
on any access of openthread_pending_dataset_obj
tulip sleet
i don't see the declaration in your repo, it is all from 4 days ago
kind river
i don't see the declaration in your repo, it is all from 4 days ago
looks like push is silently failing
kind river
so I am
tulip sleet
anyway, I'm looking at main right now. Assume you will want to reset --hard main after pushign this stuff to the branch
kind river
yes
tulip sleet
so are you saying that just typing openthread.pending_dataset into the REPL is crashing?
kind river
yes
tulip sleet
in common_hal_wifi_init(), there is this, for instance:
common_hal_wifi_radio_obj.base.type = &wifi_radio_type;
do you set the type somewhere?
kind river
That's probably it
tulip sleet
i think the pointer to the obj is fine, but stuff in the object isn't initialized
kind river
i think the pointer to the obj is fine, but stuff in the object isn't initialize...
everything should be initialized in openthread___init__
manic glacierBOT
analog pivot
I think that should go to <#537365702651150357>, but the short answer is that it...
Yeah my mistake, just noticed the "dev" part of this thread. I figured it out in the end and succeeded with my project. It was actually the USB disable mode of the macropad that I forgot existed, but a separate issue persisted and that seems to be a MacOS Sonoma thing. God, Apple sucks.
lilac scaffold
hello! I'm returning to a CP project with a new hw design.
I wanted to verify that I'm going about the design the right way - or what product has the right hw that I can look at a schematic?
I'm going to use the STM32F746 which has nothing on it AFAIK
tulip sleet
hello! I'm returning to a CP project with a new hw design.
I wanted to verify th...
There is a build: stm32f746g_discovery which uses one of STM's development boards. You could start with that. If you have the dev board that would help, since if something's not working you can compare it with that board.
many of the STM builds are rarely used, so there may be undiscovered problems
I'm not saying there are, but there might be.
lilac scaffold
that's good advice thanks
manic glacierBOT
manic glacierBOT
onyx hinge
π¦ doc building feels like it's been fragile lately
lone axle
It seems that there is a duplicate copy of the PN532 driver docs in RTD under the more general name adafruit-circuitpython https://adafruit-circuitpython.readthedocs.io/en/latest/
onyx hinge
https://readthedocs.org/projects/adafruit-circuitpython/ the maintainer of that project is not us. it was last built 4 years ago. I'll reach out to RTD via e-mail and see if they can remove it.
slender iron
that's good advice thanks
What is the board for? We have better CP support for other microcontrollers. Maybe one of them would also work.
manic glacierBOT
grizzled thunder
Issue with defining a group of Neopixels
pixels = neopixel.NeoPixel(board.NEOPIXEL, n :4) if i post this i get a syntax error
on the N: 4
The ide tells me it should exist, but it won't work.
crimson ferry
@grizzled thunder pixels = neopixel.NeoPixel(board.NEOPIXEL, 4)
stuck elbow
pixels = neopixel.NeoPixel(board.NEOPIXEL, n :4) if i post this i get a syntax e...
that's correct, this is a syntax error, don't use llms for programming
grizzled thunder
thanks, i thought i tried that too.
lone sandalBOT
kind river
Seems like my crashes are due to an openthread config issue or something because just running otInstanceInitSingle(); triggers a crash.
manic glacierBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
lone axle
Should we go ahead and make a release for the cookiecutter repo? We haven't done that in the past, and it's not required for usage during cookiecutter execution. But as changes get made over time it might be easier to keep track of them and be able to refer back to different versions by some easier to use tag number rather than back through commits or PR history.
Also if we do want to make one can anyone think of anything that it might cause issues with? To the best of my knowledge cookiecutter doesn't interact with much if anything in the nightly automated tasks and I can't think of anything else that could be impacted.
tulip sleet
Should we go ahead and make a release for the cookiecutter repo? We haven't done...
A release is just a tag. I can't think of anything that would break if it were tagged. I think this is a good idea.
lone axle
Any thoughts on what to use for the version number? I know it's arbitrary and doesn't matter much. I'm inclined to think something like 1.1.0 or 2.0.0 to try to indicate that there is some history that comes before the first numbered version.
solar whale
FYI -- I used the latest cookiecutter on my RFM repo and the transition to ruff went fairly smoothly. I had to use a few #noqa:s
lone axle
Thank you for the feedback.
tulip sleet
Any thoughts on what to use for the version number? I know it's arbitrary and do...
that is fine. You could look back at any major transitions. You could even tag some previous merge commits but probably not worth it. This is definitely a major release due to the shift to ruff
solar whale
I am getting this failure on github when I update my repo. Does anyone know what I have to fix? ```Checking dist/circuitpython_rfm-1.2.3-py3-none-any.whl: FAILED
ERROR long_description has syntax errors in markup and would not be
rendered on PyPI.
line 51: Error: Unknown directive type "todo".
.. todo:: Remove the above note if PyPI version is/will be available at
time of release.
Checking dist/circuitpython_rfm-1.2.3.tar.gz: FAILED
ERROR long_description has syntax errors in markup and would not be
rendered on PyPI.
line 51: Error: Unknown directive type "todo".
.. todo:: Remove the above note if PyPI version is/will be available at
time of release.
Error: Process completed with exit code 1.```
This is a link to the github actions failure https://github.com/jerryneedell/CircuitPython_RFM/actions/runs/10147209464/job/28057054022#step:2:1417
onyx hinge
@solar whale CI is creating files to upload to pypi. during this process, checks are made for errors in the project's "long description", and it fails because the ".. todo::" blocks are not OK in the "long description". You should resolve whatever the todo notes pertain to, and then delete them. The text it's talking about is probably in readme.rst
solar whale
<@343570430042308609> CI is creating files to upload to pypi. during this proces...
Thank you! That was it!
OK -- progress, now when I try to generate a release, I get Uploading circuitpython_rfm-0.1.1-py3-none-any.whl 25l 0% ββββββββββββββββββββββββββββββββββββββββ 0.0/38.6 kB β’ --:-- β’ ? 100% ββββββββββββββββββββββββββββββββββββββββ 38.6/38.6 kB β’ 00:00 β’ 117.3 MB/s 100% ββββββββββββββββββββββββββββββββββββββββ 38.6/38.6 kB β’ 00:00 β’ 117.3 MB/s 25hWARNING Error during upload. Retry with the --verbose option for more details. ERROR HTTPError: 403 Forbidden from https://upload.pypi.org/legacy/ Invalid or non-existent authentication information. See https://pypi.org/help/#invalid-auth for more information. Error: Process completed with exit code 1. Am I lacking some permission or am I doing something wrong?
tulip sleet
OK -- progress, now when I try to generate a release, I get ```Uploading circuit...
is this still your repo or was it transferred to the adafruit organization?
solar whale
Still mine.
tulip sleet
do you want it to be on pypi now? You need to get an accoun on pypi and supply credential info to upload from the CI. If you are willing to wait until it's transferred, it can be uploaded then
solar whale
I'm happy to wait. No rush to see it on PyPi from my perspective. I just wanted to get it ready for people to test and review.
I may have misunderstood, but in last weeks meeting, I thought you suggested that I should create a release so it can be added to the "community bundle" for review.
tulip sleet
that's the lowest overhead path for someone wanting to try it. The CI uses GitHub secrets PYPI_PASSWORD and PYPI_USERNAME to upload to pypi. We have our own secrets for our account, of course. You can add secrets to your repo under settings, or you could make an organzation and add organization secrets. Former is easier for now
solar whale
Sorry, I am missing some part of the "big picture" . So are you suggesting I create my own PyPi credentials or transfer the repo to Adafruit?
I am more than happy to transfer it.
tulip sleet
Sorry, I am missing some part of the "big picture" . So are you suggesting I cr...
we could go ahead and transfer it. You have a paid seat in our GitHub org.
tulip sleet
the name would be changed to Adafruit_CircuitPython_RFM. let me see who has to initiate it
solar whale
Ah -- OK so should I rerun Cookie Cutter to make it an Adafruit library?
tulip sleet
yes, though I'm not sure of the details that will change
you would initiate the transfer from "Settings" in the repo. Scroll down to the "Danger Zone" section on that page.
solar whale
In the past, I think someone then created a template Adadrutit library and I then pushed to it.
tulip sleet
I could do that too, since it's not a lot of stuff to copy over
would you prefer that or do you want to keep the commit history of your repo
How about I just create an empty repo and you can do a fresh cookie cutter and copy things in?
solar whale
That sounds easiest to me - I don't need to preserve the history. If you don't mind creating the template.
Hopefully Adafruit will find this worthwhile!
There is no rush on any of this.
tulip sleet
That sounds easiest to me - I don't need to preserve the history. If you don'...
ok, here you go: https://github.com/adafruit/Adafruit_CircuitPython_RFM. It's empty except for the license file.
solar whale
Great -- thank you!
manic glacierBOT
stark scroll
where is neopixel_write defined?
manic glacierBOT
It appears to me that this is resolved now. I see the content inside the board module docs page on the currently live docs. I also saw an actions run that was previously failing due to this that was able to pass now.
stark scroll
i donβt really care, iβm trying to port some stuff to Rust and Iβm looking to see how to interact with gpio21 on the adafruit feather rp2040 CAN board. The adafruit rp2040 example for rainbow doesnβt work, and iβm pretty sure itβs because the pins are different. So i just want to see the source for how to communicate with the neopixel. Either python or C would be fine. But i pulled the circuit python community bundle and the circuit python bundle and itβs not in there from what i see
slender iron
i donβt really care, iβm trying to port some stuff to Rust and Iβm looking to se...
stark scroll
ugh, itβs not a simple serial connection, thank you for this!
slender iron
ugh, itβs not a simple serial connection, thank you for this!
Nope. PIO is the best way on RP2040 (and most other micros)
solar whale
@tulip sleet I forked the new Adafruit_CircutPython_RFM library and created a brach to update it. When I try to push the branch, I am getting this... ```(blinka_venv) jerryneedell@gjnpi5nvme:~/projects/combined_rfm/Adafruit_CircuitPython_RFM $ git push origin jerryn_initial_commit
Username for 'https://github.com': jerryneedell
Password for 'https://jerryneedell@github.com':
Enumerating objects: 89, done.
Counting objects: 100% (89/89), done.
Delta compression using up to 4 threads
Compressing objects: 100% (86/86), done.
Writing objects: 100% (87/87), 64.82 KiB | 3.81 MiB/s, done.
Total 87 (delta 45), reused 0 (delta 0), pack-reused 0
remote: Resolving deltas: 100% (45/45), completed with 1 local object.
To https://github.com/jerryneedell/Adafruit_CircuitPython_RFM.git
! [remote rejected] jerryn_initial_commit -> jerryn_initial_commit (refusing to allow a Personal Access Token to create or update workflow .github/workflows/build.yml without workflow scope)
error: failed to push some refs to 'https://github.com/jerryneedell/Adafruit_CircuitPython_RFM.git'
I gave you some extra privs. Try again. If that doesn't work let me try one more priv.
lone axle
<@&356864093652516868> Sorry for the lack of additional warning today. But the weekly meeting is set to begin here on on discord in just a few moments.
solar whale
I gave you some extra privs. Try again. If that doesn't work let me try one more...
still same failure
tulip sleet
still same failure
ok, try once more
midnight ember
even though i'm here i'm text only
solar whale
ok, try once more
same failure π¦
slender iron
π»
midnight ember
π±
tulip sleet
same failure π¦
the error is "refusing to allow a Personal Access Token to create or update workflow .github/workflows/build.yml without workflow scope)". Maybe your personal access token didn't have all the right privs checked when you created it. There is a long list of privileges
solar whale
OK -- I'll try to regenerate it and try again
First hae to remeber how I created it in the first place π
midnight ember
I loaded up the Qualia S3 with 9.1.1 last night, works well.
midnight ember
takes go kart to another level. never saw that coming.
Will we get a notice of some kind when everything switches from pylint & black to ruff?
lone axle
not necessarily. Things are mostly just changing as they get worked on for other reasons.
midnight ember
looking forward to the singular RFM library. thank you for all the work on it!
All of the matter stuff looks super complicated. Glad we have you working on it. π
slender iron
Ya, its been a bunch of new stuff for me.
Complicated but I think understandably so
midnight ember
Thank you for hosting @lone axle Have a great week everyone!
I also dove into some JPG imageload this morning. Was able to cut out about 50 lines of code. It works very well.
the difference in using jpegio decoder and imageload is night and day.
lone axle
Here is the notes document for next Mondayβs CircuitPython Weekly Meeting. It is at the normal time of 11am Pacific / 2pm US Eastern here on Discord. Add your hug reports and status updates to the document before the meeting. If you are unable to attend but would still like to contribute, feel free to add your notes and weβll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/1vwgVemFzcQIq3IEBjD66LEsObJawdNTniufYQoimPdc/edit?usp=sharing
slender iron
@tulip sleet want me to review the mp merge today?
solar whale
@tulip sleet -- all set with github personal access token now. Not quite sure what was necessary to add, but it looks like I added enough....
manic glacierBOT
manic glacierBOT
manic glacierBOT
onyx hinge
ah vacation, time to work on the "posix" port again. Down to 3 failures + 97 skips, compared to unix coverage port 0 failures + 34 skips. As always, still a ways to go.
solar whale
Vacation? Sounds like a "busman's holiday" to me.
manic glacierBOT
tardy bluff
Trying to enable compiled modules on esp32 (ItsyBitsy) and I have it so that this now works, but hit
>>> import mandelbrot
ets Jul 29 2019 12:21:46
rst:0xc (SW_CPU_RESET),boot:0x13 (SPI_FAST_FLASH_BOOT)
configsip: 271414342, SPIWP:0xee
clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00
mode:DIO, clock div:1
load:0x3fff0030,len:396
ho 0 tail 12 room 4
load:0x40078000,len:13916
load:0x40080400,len:4
ho 8 tail 4 room 4
load:0x40080404,len:3156
entry 0x40080558
Serial console setup
Is there any guidance for how to debug this? Context:
https://github.com/adafruit/circuitpython/issues/9471
Appreciate that this is a low priority for the adafruit team so keen to try and get this working myself
manic glacierBOT
manic glacierBOT
manic glacierBOT
slender iron
@tardy bluff Do a debug build and look at the UART output. It should have a backtrace. I bet that reset is due to a call to reset_into_safe_mode()
tardy bluff
<@1267746759124320267> Do a debug build and look at the UART output. It should h...
Thanks for the pointer, will have a look at that in a bit (will need to solder some pins on π€¦ββοΈ) but good to know that is where the output goes, appreciated
slender iron
thanks for looking into compiled modules! Others will be excited to have it
lone axle
Does anyone have any recent experience with cookiecutter on windows? I didn't realize there was a different branch of the cookiecutter template and instructions for using ansys-cookiecutter instead of "regular" cookiecutter noted here https://learn.adafruit.com/creating-and-sharing-a-circuitpython-library/creating-a-library#running-cookiecutter-windows-3114568.
It looks like this windows branch is out of date though and a user other in #help-with-circuitpython is having some trouble trying to use it on Windows.
I'm wondering if there is some way we could make the process easier on windows or if there is interest in doing so. Could we maybe have some Github repo which uses specialized actions to run cookiecutter and then make the resulting project zip available for download? Kind of like how we have a way to make a custom build of core circuitpython via github actions?
If we could change that process from command line tool requiring a few libraries and other utilities to be installed over to just a "form" page that allows the user to enter their answers into standard inputs on a webpage it could help lower the friction for people wanting to share their libraries.
It also turns out that ansys-cookiecutter which the guide recomends doesn't seem to exist any longer. https://pypi.org/search/?q=ansys-cookiecutter returns no results by that example name and pip install ansys-cookiecutter fails with the error No matching distribution found for ansys-cookiecutter
Re: using github actions to run cookiecutter. Looks like this has at least been done before https://medium.com/@jimmfan/automation-with-cookiecutter-and-github-actions-f7dd8f7f042e
manic glacierBOT
candid sun
Does anyone have any recent experience with cookiecutter on windows? I didn't r...
i'm on windows but use linux subsystem whenever i run cookiecutter
manic glacierBOT
orchid basinBOT
onyx hinge
@lone axle cookie cutter is now tested during ci that it works (that is, runs to completion) on Windows. The need for a fork is no longer true.
manic glacierBOT
lone axle
I created a proof of concept server app that basically converts cookiecutter questions into a web form instead of CLI https://cookiecutter.foamyguy.com/cookie/?template=gh:adafruit/cookiecutter-adafruit-circuitpython. It appears to be working successfully but I haven't done exhaustive testing or validated everything inside of the generated projects. I think something like this might be a little more approachable to some folks than the existing CLI cookiecutter process. Theoretically it could handle other cookiecutter templates by changing the URL param, but I've limited it with an allow list to only work for this one ATM. Because I kept it general enough to work with any cookiecutter the "hints" given in the form page are kind of cryptic since they're just the default values, if you don't understand jinja template and cookiecutter they may not mean anything to you. There could be a form made more specific to circuitpython cookiecutter though with more specific and helpful hints if this is something that we think might be a useful tool.
Moving this into a server has the benefit of taking the burden of getting python and cookiecutter CLI setup out of the end users hands and into a central server that anyone can use from any OS with only a browser and no additional setup. Initially part of my motivation was the differing process on Windows, but I understand now there is no longer a need for that different process, though we do still need to update the learn guide to reflect it.
Currently it's a Django app because that is what I'm most familiar with and could throw together quickly. I do think it might be possible to utilize github actions to move the generator "server" into workflow runners instead of an actual server deployed and accessed via the browser. However witht he github actions method I don't know about the possibility of making simplified all the way down to a form like this. The process might end up still being something like: fork/Clone the "cookiecutter wizard actions repo", edit a JSON config file with the answers to your cookiecutter questions, then commit / push and the actions workflow can grab your config file and run cookiecutter with it then upload the resulting zip into the repo, or maybe elsewhere. "The barrier to entry" is maybe lower than cookiecutter CLI directly, but not by very much I think.
manic glacierBOT
Signal quality is never going to be good with the RP2040 ADC so I'm shifting focus to an I2S audio codec, TAC5212. I've got the ADC and DAC working and am working on implementing I2SIn a la #5456.
orchid basinBOT
shadow vigil
hmmm - i'm having that crazy [expletive deleted] pre-commit issue on a CP library: i'm using Ubuntu 24.04, Python 3.12 and I've tried using "fresh" virtualenv and venv without any success (plus the "well did you install pip" stuff) -- any clues or hints?
spare jacinth
i can try later on an Ubuntu24 WSL image, for the sake of reproducing it:
python -Vshows 3.12- then, you make a new env with
python -m venv some_name - then
. some_name/bin/activate(orsourceinstead of.) - then
pip install pre-commit - then
pre-commit installfails? or is it an error thrown when actually running the checks (when doinggit commit)?
shadow vigil
yes - AttributeError: module 'pkgutil' has no attribute 'ImpImporter'. Did you mean: 'zipimporter'? which is all over my search but i've not found an actual working solution
spare jacinth
- fresh image, ran
apt update && apt upgrade apt install python3(which in fact is 3.12)- then, trying to run
python3 -m venv venvfails with a hint message apt install python3.12-venv(as the hint suggested)- now we can actually
python3 -m venv venv+. venv/bin/activate pip install pre-commitworks π€
shadow vigil
python3.12-venv/noble-updates,noble-security,now 3.12.3-1ubuntu0.1 amd64 [installed,automatic]
pre-commit -V pre-commit 3.8.0
spare jacinth
so it seems to be installed (pip install pre-commit worked)
does it fail when you try to configure it on the repo (pre-commit install)?
when you try and commit and the hook gets to run?
shadow vigil
yes, it fails with the above attribute error on run
spare jacinth
i asked whether it fails with a command or another, "yes" doesnt answer that π€£
if it fails when running the hooks on your .pre-commit-config.yaml, maybe they are not compatible with 3.12, and is not a pre-commit issue on itself
shadow vigil
it's failing on run with the error above -- and it's the Blinka library, so not a new one -- i'm re-trying to do my change on a Raspberry Pi that has 3.11
(it's so stupid simple i might just give up and do it on github itself π )
lone axle
yes - `AttributeError: module 'pkgutil' has no attribute 'ImpImporter'. Did you ...
I think this is an error related to python 3.12 and some specific versions of pylint that we had pinned that get used by pre-commit. As a temporary workaround you can update the pylint version to 3.x locally and then run your pre-commit successfully. In the longer term we're working on moving to Ruff instead of pylint which will also mean this problem goes away.
lone axle
For the workaround, inside of .pre-commit-config.yaml you can change the version on the 2nd of these lines:
- repo: https://github.com/pycqa/pylint
rev: v2.17.4
to
rev: v3.0.4
so yeah, it was a problem with some code on the hooks that pre-commit was being told to use, not something within pre-commit itself
shadow vigil
yes, it's the pylint version -- i can do a PR for this one, too π
manic glacierBOT
vital reef
Hey all! I'm working on the ESP32 light sleep project, and kind of hitting a wall here. It just doesn't seem like it's possible in any nice/satisfying way on most boards.
Light sleep is supposed to keep everything running in the background, but there's no way to tell if you're connected to USB if it's through a USB to serial chip.
There's also no way to receive serial data while in sleep mode as far as I can tell. You can wake on serial, but the first character is lost.
I suspect there's plenty of other edge cases like this too. We could do some kind of "Assume connected until X seconds after the last char" but that's not completely reliable.
Is there something I'm not seeing here? It seems like making effective use of sleep just isn't possible without some kind of compromise to the existing APIs, new flags to disable certain features during sleep, or software implementations of stuff on the ULP cores.
stuck elbow
I'm trying to implement rgb png images for imageload, and when I try to create a 24-bit displayio.Bitmap, I get ValueError: value_count must be 1-65536 β am I supposed to somehow be converting the image data to RGB656 on the fly or something?
Also, which colorspace should I use for grayscale images?
stark aurora
Is there Nordic NRF Bluetooth support ?
stuck elbow
Is there Nordic NRF Bluetooth support ?
what are you trying to do?
stark aurora
Frequency counter counts frequency 20 or 50 times/sec and PWM peripheral drives LCD lens. Controller communicates 9600 bps with external device to set PWM as function of frequency deviation (diff. to se table ref. frequency).
Iβm not sure using AI in Forth could solve/learn relationship between frequency difference and required PWM value starting with some initial settings? If this would be possible this could be a next candidate for smallest FPGA ICE40 device of new 130nm CMOS ASIC
stuck elbow
it's hard to beat an fpga
stark aurora
Size and power is the issue here⦠With ICE40 relative low power is possible but load config by startup from onboard flash consumes much more energy then in sleep/action modes
stuck elbow
circuitpython really isn't designed for low power or efficiency
you are probably going to get best results with low-level languages
stark aurora
Have done it already with less than 256 lines assembly code
Maybe HANA1 cpu [ Christopher Lozinski ]
stuck elbow
the only nrf mcu supported at the moment is nrf52840, it's pretty big and beefy
that's because circuitpython needs a lot of ram
stark aurora
Are there some examples. Sending some parameters for the embedded control loop and receiving embedded measurement data to/from mobile Android/IOS device ?
stuck elbow
Are there some examples. Sending some parameters for the embedded control loop a...
something like this? https://learn.adafruit.com/circuitpython-nrf52840/mobile-movement-data
you probably want to go to #help-with-circuitpython, there might be more people able to help you there, this channel is mostly for development of the circuitpython and its libraries
or even better, in #help-with-projects people might be able to suggest other solutions than circuitpython
stark aurora
No support yet for nrf5340 dual core ? Would like to use one core for Bluetooth and the other as control loop handling some peripherals. Or nrf52840 but then scheduling is required I believe and donβt know if it is supported yet in circuit Python
stark aurora
Running Bluetooth and embedded control loop at βthe same timeβ
stuck elbow
if anybody is familiar with the font bitmap data generation, I would love some help with https://github.com/adafruit/circuitpython/pull/9479
midnight ember
if anybody is familiar with the font bitmap data generation, I would love some h...
Are you talking about the bitmap font for terminalio and blinka logo for repl?
onyx hinge
Hello! No circuitpthon work to report. Instead, I'm on safari. Group hug!
lone sandalBOT
stuck elbow
Hmm, it looks like the "windows" job is failing with the "This environment is externally managed" error. Did Python get updated?
spare jacinth
I dont recall seeing that on my Win computer, and i use 3.12, which afaik is the latest stable release
stuck elbow
it happens in recent versions of Python when there is a special file in /etc that tells people to install stuff witht he package manager and not pip
spare jacinth
Oh, ok
tulip sleet
Hmm, it looks like the "windows" job is failing with the "This environment is ex...
I noticed the windows job failure in https://github.com/adafruit/circuitpython/pull/9469 and have a fix there. This job vets that building on Windows still works.
stuck elbow
files changed: 362 π±
tulip sleet
that is less than the usual merge from upstream π
stuck elbow
should I include the fix in my patch, or just wait for that patch to be merged?
or maybe a separate patch with just the fix?
tulip sleet
should I include the fix in my patch, or just wait for that patch to be merged?
we will just ignore the failing windows jobs when considering whether to merge PR's right now. That job is not on any critical path.
stuck elbow
ah, ok
tulip sleet
i'm not sure if there's a better way to do this, but setting a venv and using it in different build steps does not seem easy to me.
stuck elbow
yeah, I remember the pep that introduced that functionality
tulip sleet
see the comment there. I researched this a bit and it seemed like the easiest way. Since msys2 does not depend on a "system" python as far as I know, this seems ok (unlike, say, doing that on an ubuntu base system)
stuck elbow
it's a virtual machine anyways, so you don't care about mangling system packages
I mean, single-purpose virtual machine