I forget why I thought I wanted the jmp pin, I think it was maybe just for completeness
#circuitpython-dev
manic glacierBOT
slender iron
@tulip sleet you around? I'm just about free
tulip sleet
yes, and I have more info
slender iron
kk, will let you know when I'm free
onyx hinge
our progress bar is getting so full
onyx hinge
well that is weird. ```$ ./micropython-coverage -c 'import sys as errno; print(errno.EINVAL)'
22
print(errno.EINVAL) even prints a number without importing anything ๐ค
tulip sleet
looks like some short-circuiting of a lookup??
onyx hinge
yeah
I understand why, I wouldn't have discovered it otherwise
but the more exciting news is, I found 1k of flash space in the couch cushion while looking
tulip sleet
lol
tulip sleet
i would think we need it =1?
onyx hinge
what makes you say that? It doesn't seem like referring to errno constants is a performance critical path in the type of code our users create
tulip sleet
// Whether to enable lookup of constants in modules; eg module.CONST . Is that an optimization or it just doesn't work if off?
It sounded like the latter
onyx hinge
oh I see what you're asking
I think it's just an optimization
.. turning errno.EINVAL into a constant load whereever it appears, rather than actually looking it up in the errno module by string at runtime
I'll make sure though
onyx hinge
disabling MICROPY_COMP_MODULE_CONST is +112 bytes free, disabling MICROPY_COMP_CONST_FOLDING is +884 for a total of +996
tulip sleet
const folding is pretty important, i wouldn't turn that off
onyx hinge
I won't PR it right now but like some other things I'll keep it in my pocket. We might decide it's a trade-off to make for the most constrained boards. constant folding would remain active in mpy-cross so in byte-compiled files it's not giving anything up
while True: print(1+2+3+4+5+6+7+8+9+10) I imagine it can make code like this run a lot faster/slower, for sure
tulip sleet
especially floating point calcs on non-fp chips
onyx hinge
yeah I see that
slender iron
ok @tulip sleet
manic glacierBOT
manic glacierBOT
spiral elk
I have some questions regarding submitting a new board if someone has some time to explain.
analog bridge
I have some questions regarding submitting a new board if someone has some time ...
have you taken a look at https://learn.adafruit.com/how-to-add-a-new-board-to-circuitpython/overview
spiral elk
Yes, the thing is that the board I'm working with is a variant of a board adafruit sells, but wondering what to do about usb pid/vid
This is more a political thing than technical.
analog bridge
You will need new vid/pid... if your board is open source then you can get it from https://pid.codes/
spiral elk
Not an open hardware board, same physical pcb as an adafruit-sold third-party product but a different mcu onboard.
manic glacierBOT
tulip sleet
@idle owl if you have availability today let's talk about keypad/gamepad stuff
idle owl
<@!330227457296957440> if you have availability today let's talk about keypad/ga...
I do. By the time I was free yesterday, I believe you were chatting with Scott, so I didn't bother you. I'm free after 11:30 it looks like. (Shipping a guide right now.) Does that work for you?
tulip sleet
that's good
idle owl
Excellent. I'll ping you in a bit.
manic glacierBOT
slender iron
Not an open hardware board, same physical pcb as an adafruit-sold third-party pr...
You'll need your own vid/pid pair. this FAQ has entry on where you can get a PID from the mcu vendor: https://pid.codes/faq/
idle owl
manic glacierBOT
We should revisit this. As @shuyuan-liu points out, the Pico datasheet designates defaults now:
lone axle
@idle owl I noticed the PR you linked earlier that removed were_pressed from CircuitPlayground library. The CLUE library contains the same API currently. I'm about to update that library for these changes. Should I remove it from there as well? I found two uses of it in examples / learn guides so if we do remove I'll update those to use the direct button value API but keep the same behavior from users perspective.
idle owl
<@!330227457296957440> I noticed the PR you linked earlier that removed `were_pr...
One of the examples in Learn isn't relevant, because it's implementing a few things from the CLUE library in a library included with the project bundle (Dan and I were looking into this earlier). Yes, we're trying to slowly deprecate gamepad and the were_pressed functionality isn't really necessary. Folks can use keypad for similar if desired, or poll the buttons etc. So if you're already in there, please remove it, and ensure that any example code is updated. Thanks!
lone axle
@tulip sleet do you know if there is a "simpletest" type most basic example of how to use the new keypad module somewhere? I've just written this one to make sure that I understand how it works:
from keypad import Keys
import board
import time
pins = [board.BUTTON_A, board.BUTTON_B]
keys = Keys(pins, value_when_pressed=False, pull=True)
while True:
if keys.events.overflowed:
keys.events.clear()
event = keys.events.get()
if event:
if event.pressed:
print("{} was pressed".format(event.key_number))
if event.released:
print("{} was released".format(event.key_number))
If there isn't one out there yet I'd be happy to share this where is the best place to do so (in the docs perhaps?)
tulip sleet
<@!329766224093249548> do you know if there is a "simpletest" type most basic ex...
I wrote this guide: https://learn.adafruit.com/key-pad-matrix-scanning-in-circuitpython
which has several simple examples
lone axle
Ah, thank you. didn't think to check for a guide. ๐คฆ
tulip sleet
np! I try to talk it up ๐
@lone axle the "overflow" stuff is discussed on the Advanced Features page
lone axle
Nice, thank you
tulip sleet
@slender iron verrrrry interesting re audio:
- stopping neopixel statemachine on deinit made no difference
- playing a
RawSampleby reading a raw version of my a440.wav file works fine by itself, even with displayio DMA going on. - playing the
RawSampleand theWavFileversion are very different. The WavFile version is crackly or very quiet. After some number of plays the RawSample play also stops working (perhaps due to DMA getting messed up).
This is with a 32k 1-second 16-bit sample. I will continue to look at things.
spiral elk
@slender iron Should I actually move ahead and prep a PR for the changes necessary to reclaim the nvm area for the blackpill_with_flash? My local build without it seems fine after a day in use.
thorny jay
Hi, I tested adafruit-circuitpython-circuitplayground_bluefruit-en_US-20210723-5fbe6d4.uf2 and noticed that boot_out.txt is empty. I tried import storage; storage.erase_filesystem() but still empty.
spiral elk
Hi, I tested `adafruit-circuitpython-circuitplayground_bluefruit-en_US-20210723-...
I can verify this on a build I did yesterday:
Adafruit CircuitPython 7.0.0-alpha.5-22-ge790ff732-dirty on 2021-07-23; stm32f411ce-blackpill-with-flash with STM32F411CE
thorny jay
I can verify this on a build I did yesterday:
`Adafruit CircuitPython 7.0.0-alp...
I tested the same 5fbe6d4 on Metro-M4 and my boot_out.txt is OK: Adafruit CircuitPython 7.0.0-alpha.5-32-g5fbe6d480 on 2021-07-23; Adafruit Metro M4 Airlift Lite with samd51j19 so I would not worry too much, there might be a glitch specific to CPB or nRF or maybe it is just me.
manic glacierBOT
jaunty juniper
hey, @tulip sleet I don't know if you can help with this, but it relates to pyserial for windows: on win 10 at least it gives me bad information for the manufacturer ("microsoft" for a CPX) and description (it says a generic "USB serial peripheral (COM3)") and doesn't even fill product (which should say "CircuitPlayground Express").
I think I understand that you used a DEVPROPKEY to get the serial channels names, do you know if a similar method would work to retrieve those other fields ? and how would I go about finding the key for those USB properties (if any) ?
ornate breach
The addition to support multiple status LEDs, will that end up working if Iโve got a matrix of LED connected for the neopixel status?
For instance the RGB matrix board I showed off on show and tell has an ESP32-s2 and currently it just uses the first neopixel of the matrix for status
Would I be able to use all the neopixel for status?
onyx hinge
Seems like?
devout jolt
Does anyone here have an Arduino Nano RP2040 they can test CircuitPython on? In the help-with-CircuitPython channel there seems to be recent issues with this board, where MicroPython and Arduino UF2 uploads work but CP does not. Sounds kinda like flash setting not right in the build?
jaunty juniper
mine works as expected, demo code runs on latest, wifi, RGB LED, I'm wondering if there are different flash chips or some variability (or something else entirely)
devout jolt
Thatโs what Iโm wondering as well. Whereโs that Arduino sketch that prints out Flash info ๐
tulip sleet
hey, <@!329766224093249548> I don't know if you can help with this, but it relat...
I am looking for the docs.microsoft.com webpages I used. Here is one for DEVPKEY_Device_BusReportedDeviceDesc: https://docs.microsoft.com/en-us/windows-hardware/drivers/install/devpkey-device-busreporteddevicedesc. Its siblings may be what you want. I am looking for the GUID values: I can't remember whether I got those from a header file or a webpage.
I think you know about this PR I made: https://github.com/pyserial/pyserial/pull/571/files
Google 540b947e-8b40-45bc-a8a2-6a0b894cbda2 and you can find the docs.microsoft.com webpages, and then look at their siblings.
(I had that GUID in my search history)
i am no expert at this: I just replicated how the existing .py script looked up values, and probably found how it was done in some other examples (maybe C#)
jaunty juniper
ah thanks a lot, looking up the keys in that format does return resources with lists and stuff, that's my new rabbit hole to dive in !
onyx hinge
I don't know whether this is a good idea, but I learned a lot to implement it.
manic glacierBOT
It's worth noting that upstream lists only one flash chip at https://docs.arduino.cc/hardware/nano-rp2040-connect
Anybody know how to properly reach out to Arduino about this?
This is the image from discord that someone was able to determine would work with the IS25LP128F flash definition
manic glacierBOT
manic glacierBOT
I think the pin assignment for the CAN pins in the Feather STM32F405 Express board are swapped.
solar whale
Sorry, I did not get to this and won't be able to test it for the next week since I am away from my hardware.
thorny jay
Sorry, I did not get to this and won't be able to test it for the next week sinc...
Dunno exactly what I tested, but with @lone axle during the live show we tested the same thing and the problem was only on my side... Like he had the boot_out.txt and mine was empty. So it might be me or I had another build.
spiral elk
Dunno exactly what I tested, but with <@382939733107408897> during the live sho...
There's clearly something wrong with some builds though and it spans multiple devices.
stuck elbow
@onyx hinge I've stolen your example, and I have to report that it works very well! https://hackaday.io/project/174095-dorsch-40k-keyboard/log/195783-shiny-new-code
onyx hinge
<@!426200547024961539> I've stolen your example, and I have to report that it wo...
very nice, hope lots of folks pick up on this stuff!
onyx hinge
@stuck elbow Is the check for "code == 0" an optimization? otherwise, it seems unneeded, since a value of 0 wouldn't pass into either of the other if-clauses. python for code in pressed_keys: if code == 0: continue if code & 0xff00: report_mod_keys[0] |= (code & 0xff00) >> 8 if code & 0x00ff: report_bitmap[code >> 3] |= 1 << (code & 0x7)
lone axle
@tulip sleet was gamepad removed in 7.0.0 alpha 5 but going to be added back for next alpha / beta? If I understand correctly the PR removing that hasn't been merged yet, but testing out something unrelated I'm getting a Module Not Found on a CLUE with alpha 5:
Adafruit CircuitPython 7.0.0-alpha.5 on 2021-07-21; Adafruit CLUE nRF52840 Express with nRF52840
>>> import gamepad
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
ImportError: no module named 'gamepad'
We are going to put it back for 7.0.0 or fix the relevant libraries. If will definitely be gone for 8.0.0. Probably we'll just put it back.
It is turned off by default in circuitpy_mpconfig.mk
@lone axle I am now seeing your other PR's. We can remove were_pressed from the CLUE library, for instance, as you did. We just have to make it a major version bump , as Kattni just did for the circuitplayground library.
What were you testing that ran into needing gamepad on the nRF?
lone axle
<@!382939733107408897> I am now seeing your other PR's. We can remove were_press...
I am working on reviewing the PRs in the Learn Guide repo that remove max_size parameter from displayio.Group() specifically this one is the one I noticed it with: https://github.com/adafruit/Adafruit_Learning_System_Guides/pull/1638 it's using the CLUE library
tulip sleet
ok, thanks. I see some uses of gamepadshift and gamepad in various learn guides, but not were_pressed(). There is were_pressed() in Pyloton, but that's a special case, and we know about that (it has its own mini version of the CLUE library). gamepadshift is still present in the boards that need it. So in the next alpha, I will put back gamepad.
@lone axle do you want the Learn Guide fixes you approved to be merged, or do they need to be done in conjunction with any guide editing?
i am just wondering why you didn't merge immediately.
lone axle
<@382939733107408897> do you want the Learn Guide fixes you approved to be merge...
I think they can be merged. I have access to merge them actually I can do that as I work through them as well. But didn't know if anyone else needed consulting before doing so
I'll start merging them as I go through. And I will update the guide pages as needed for any that have code outside of github embeds.
tulip sleet
@onyx hinge I am seeing this in local make html builds. Is it significant?
checking consistency... /home/halbert/repos/circuitpython/shared-bindings/ulab/numpy/approx/index.rst: WARNING: document isn't included in any toctree
/home/halbert/repos/circuitpython/shared-bindings/ulab/numpy/numerical/index.rst: WARNING: document isn't included in any toctree
/home/halbert/repos/circuitpython/shared-bindings/ulab/numpy/stats/index.rst: WARNING: document isn't included in any toctree
/home/halbert/repos/circuitpython/shared-bindings/ulab/numpy/transform/index.rst: WARNING: document isn't included in any toctree
/home/halbert/repos/circuitpython/shared-bindings/ulab/numpy/vector/index.rst: WARNING: document isn't included in any toctree
build finished with problems, 5 warnings.
make: *** [Makefile:93: html] Error 1
FYI
halbert@tuna:~/repos/circuitpython/extmod/ulab$ git status
HEAD detached at 8d93dde
nothing to commit, working tree clean
@tulip sleet that's interesting, can you do a git clean in shared-bindings and try it again? Is it failing on github too, or just locally for you?
I think numpy.numerical is a module that has been removed/renamed into numpy-compatible naming
but I notice that building the docs leaves .rst files in shared-bindings/ that are maybe being picked up later?
tulip sleet
i was just testing in preparation for a PR for gamepad-related stuff. About to go for a walk but I will try again in about an ahour
tulip sleet
<@!329766224093249548> that's interesting, can you do a `git clean` in shared-bi...
a thorough git clean did the trick. Thanks!
manic glacierBOT
slender iron
<@!252717193496756235> Should I actually move ahead and prep a PR for the change...
Sure!
manic glacierBOT
stiff pelican
Dunno exactly what I tested, but with <@382939733107408897> during the live sho...
@solar whale I'm getting a blank boot_out.txt as well running a S3-latest on Pimoroni Pico Lipo 16MB
thorny jay
<@!343570430042308609> I'm getting a blank `boot_out.txt` as well running a S3-l...
I was thinking it was just a glitch on my side... There is no issue created yet, but I am at work, so I prefer not to access GitHub during working hours. Maybe you can create an issue for that empty boot_out.txt?
I am always hesitant to create an issue... I prefer to check here if it is a know behaviour, if I am the only one affected, before creating an issue. You see, I don't track the forum, I might only see a small percentage of the created issue. So I almost never know for sure if it is a known issue, a new issue, just me making mistake or not using new code the expected way.
manic glacierBOT
manic glacierBOT
turbid radish
?serverinfo
digital shoreBOT
adafruit
tulip sleet
@idle owl gamepad + keypad together is not fitting in Feather M0, etc. so I may just remove gamepad altogether from SAMD21, now that the libraries have stopped using it. It does fit on CPX because we turn off displayio, which is big
idle owl
<@!330227457296957440> gamepad + keypad together is not fitting in Feather M0, e...
So it won't fit on the displayio build then.
I'm all for turning it off.
tulip sleet
we pull other tricks on the displayio build (like no longints)
idle owl
Ah
tulip sleet
i will fix pyloton not to use it
idle owl
i will fix pyloton not to use it
Sounds good.
<@&356864093652516868> Greetings! The CircuitPython Weekly meeting is in just under 2.5 hours. Remember that if you'd like to participate (or are missing the meeting and wish to provide notes), please add your hug reports and status updates to the notes document. This is what I will pull from as I go through the meeting. If you're planning to simply listen in, you don't need to do anything. Hope to chat with you soon! https://docs.google.com/document/d/1IO_iC2RE9w83MMxA-Gy5T3VeeAy3HbuCkSOehqth15I/edit#
jaunty juniper
somebody in the help channel had installed the latest libraries from the bundle on a CPB 6.3.0, which is still the stable, and of course it didn't work because of keypad. The issue here is that the problem is not obvious: the bundle is described as "This bundle is built for use with CircuitPython 6.x.x." and it's actually not, because that's about the mpy format used, not the feature compatibility
the solution, to download an older version of the adafruit_circuitplayground library from the github releases of said repository is not super easy to find on your own either
thorny jay
somebody in the help channel had installed the latest libraries from the bundle ...
There were some discussion with @lone axle on how to handle the migration. But it seems that now we keep both... so the library can be changed later... after 7 is released.
jaunty juniper
yeah but the library is already changed, has not backwards compatibility and is in the bundle
lone axle
somebody in the help channel had installed the latest libraries from the bundle ...
Did they post the specific error message? If I understand correctly the usage of keypad was also removed so I would have thought that library would work with 6.3.0
jaunty juniper
oh it was ?
jaunty juniper
ah ok, I see
lone axle
the were_pressed API was removed all together with this PR. So user code can either poll button_a or button_b directly or if desired implement keypad in the usercode space instead of within the library.
jaunty juniper
so it was in 4.4.0, but is fixed in 5.0.0
idle owl
We broke it temporarily.
In terms of 6 compatibility.
It has been fixed in the latest version, yes.
jaunty juniper
ok good, thanks ๐
idle owl
No one used were_pressed anyway, and shouldn't have been. That's all that ended up changed in the end was its removal.
tulip sleet
were_pressed() is also going to get removed from the CLUE library: https://github.com/adafruit/Adafruit_CircuitPython_CLUE/pull/48
jaunty juniper
well I use it but that's fine ๐
tulip sleet
you are "encouraged" to switch to keypad ๐
idle owl
well I use it but that's fine ๐
keypad is better ๐
thorny jay
Obviously you are better informed and I have a partial view. Long LIVE the keypad.
idle owl
Personal opinion, I suppose.
thorny jay
I work in "operation" (the Ops in DevOps) and I was: "But how are we going to transition from the old to the new and synchronise the library change and the core change."
tulip sleet
I work in "operation" (the Ops in DevOps) and I was: "But how are we going to tr...
Ideally we'd leave gamepad in for one more version, but there are space issues, as you see. So by trimming the libraries a bit, we can force the change. It's not quite as pleasant as an overlapping transition, but it's necessary in this case.
The frozen CPX library is OK for now in 6.3.0
that will not change in 6.x
jaunty juniper
yeah, that's why their code worked on the CPX but not the CPB
analog bridge
@slender iron I'd like to discuss about the traceback module implementation, can you ping me when you are available.
tulip sleet
in the long run we might have different bundles, but we haven't come up with an easy mechanism for that.
there's a long issue discussing that
slender iron
@analog bridge at my desk now
analog bridge
I have implemented the cpython version of print_exception(), the only thing is that the cpy exceptions doesn't contain the __traceback__ obj...
Should I alter the print_exception() function to only take exception (i.e. print_exception(value), instead of print_exception(etype, value, traceback))?
the changes are here: https://github.com/microDev1/circuitpython/tree/traceback
cpython implementation: https://docs.python.org/3/library/traceback.html
slender iron
could we add __traceback__?
another option is to make a new module that does it based on the exceptionn
and then we can make a python version for cpython
analog bridge
yes the mechanics are there... its just not exposed
jaunty juniper
the C python traceback API is not great, I personally would prefer we implemented traceback.format_exc which can only be called for the latest exception, but the result can be kept for later so it's fine, and it doesn't require additional parameters
or, micropython-lib has a traceback module written in python
analog bridge
the C python traceback API is not great, I personally would prefer we implemente...
good idea... one can just print traceback using print(traceback.format_excecption())
jaunty juniper
the only thing is I guess it allocates memory
tulip sleet
there is an emergency exception buffer, could that be used?
analog bridge
or, micropython-lib has a `traceback` module written in python
seems like it doesn't provide any traceback module features like limit, chain
https://github.com/micropython/micropython-lib/blob/3a6ab0b46d6471bee00ae815444c624709dd4cdd/python-stdlib/traceback/traceback.py
tulip sleet
MICROPY_ENABLE_EMERGENCY_EXCEPTION_BUF, etc.
jaunty juniper
seems like it doesn't provide any `traceback` module features like `limit, chain...
yeah no I was thinking backwards: if it's in a new library the compatibility solution is what Scott says, have a python implementation in blinka, not have a traceback.py emulation in Circuitpython
slender iron
I think there are two options 1) add native traceback that is a subset of CPython's traceback or 2) create a new module name that has the existing print_exception and provide a Python version in Blinka for CPython. Either option is fine with me. My goal was to get it out of sys
analog bridge
could we add `__traceback__`?
following seems to be the place were __traceback__ implementation will go?
https://github.com/adafruit/circuitpython/blob/main/py/objexcept.c#L207-L253
slender iron
yup, looks right
analog bridge
one more thing... we don't actually need to obtain etype, traceback params from the user... they can be obtained in c from the exception i.e. value parameter.
in-fact python-3.5 just ignores etype, its only there for backwards compatibility with things like sys.exc_info() which returns a tuple (etype, value, traceback), exc_info() is at the moment disabled by default, should I enable it?
analog bridge
I think there are two options 1) add native `traceback` that is a subset of CPyt...
my concern is with flash usage... I an not sure if traceback will fit on all boards.
slender iron
we don't need it everywhere
analog bridge
I am removing etype, traceback params for now and will focus on adding print_excepction(value, limit=None, file=None, chain=True)
and also format_exception() its same as print... just returns a string instead printing or writing to file like object
slender iron
I am removing `etype, traceback` params for now and will focus on adding `print_...
if the calls won't work in CPython please call it something besides traceback
analog bridge
need help with docs... is it possible to have params without type annotations?
//| def print_exception(etype, value: BaseException, tb, limit: Optional[int] = None, file: Optional[io.FileIO] = None, chain: Optional[bool] = True) -> None:
//| """Prints exception information and stack trace entries.
//|
//| .. note: Setting `chain` will have no effect as chained exceptions are not implemented.
//|
//| :param etype: This is ignored and inferred from the type of value.
//| :param BaseException value: The exception.
//| :param tb: The traceback object.
I am struggling with etype and tb...
slender iron
There is an Any type: https://docs.python.org/3/library/typing.html#typing.Any
idle owl
I will be right back.
modern wing
I will be right back.
thanks for the warning ๐
storm minnow
I wish I could join the frey
modern wing
unplug the second monitor so it doesn't randomly reboot?
thorny jay
I will follow the recording. I wrote my notes (that cover some hug report from last week too).
onyx hinge
Meeting calendar for adding to your calendar service: https://raw.githubusercontent.com/adafruit/adafruit-circuitpython-weekly-meeting/master/meeting.ical
For viewing online: https://open-web-calendar.herokuapp.com/calendar.html?url=https%3A%2F%2Fraw.githubusercontent.com%2Fadafruit%2Fadafruit-circuitpython-weekly-meeting%2Fmaster%2Fmeeting.ical&title=CicuitPython Weekly Meeting Schedule&tab=agenda&tabs=month&tabs=agenda
onyx hinge
open horse hardware ๐
blissful pollen
I needed to light up my life...
onyx hinge
that's a lot of unfamiliar names
modern wing
๐ต So please carry the lantern high. ๐ก
onyx hinge
Add Gravitech Cucumber R board definition (updated) #4131
storm minnow
https://twitter.com/todbot/status/1417297915094077450
Haha, glad to see someone was able to successfully do something "wrong" with a Trinkey after my fiasco with the defective Slider Trinkey mod
Good job @devout jolt
swift arrow
very sorry, didn;t notice I wasn;t
slender iron
np
lone axle
maybe the script is checking for (non-pre-) releases?
slender iron
ahead of the game as always ๐
idle wharf
@slender iron Deep Dive question.... is there a page where we can see all the notes\timestamps across all the previous dives? I wanted to re-watch some sessions where you're doing protocol analysis
slender iron
@idle wharf I don't have a single place. We could make a repo for it like we do for these meetings
slender iron
I can probably jam it out
modern wing
Thanks everyone ๐
onyx hinge
๐
slender iron
Let me know how I can help make it happen
blissful pollen
Just a small FYI real life has really got in the way of contributing much, but always feel free to message me if there is a question on something I worked on
slender iron
np @blissful pollen. I hope all is well
blissful pollen
np <@!228322543507865601>. I hope all is well
Yup, nothing major just work, covid stress, and the like. But things are looking better soon
slender iron
Yup, nothing major just work, covid stress, and the like. But things are looking...
๐ take care of yourself. let us know if you need help with anything
blissful pollen
Thanks! All this is a great distraction and fun to pay attention to
idle wharf
https://github.com/adafruit/deep-dive-notes
Are the older notes docs available or should I work on a script to scrape YT tonight ?
slender iron
I'm adding them from the google drive copies
tulip sleet
@idle owl I am adding board.LED to those boards that should have it. But I see a bunch of boards with board.L: should I leave that or change it to board.LED?
idle owl
<@!330227457296957440> I am adding board.LED to those boards that should have it...
Leave it. Don't remove anything already present.
idle owl
<@!330227457296957440> I am adding board.LED to those boards that should have it...
Never know where it was used, and don't want to break things.
tulip sleet
โ๏ธ will just alias
idle owl
Perfect.
crimson basin
<@&356864093652516868> Greetings! The CircuitPython Weekly meeting is in just un...
I would like to attend the meeting. This would be the first time. How can I join this meeting?
idle owl
I would like to attend the meeting. This would be the first time. How can I join...
Today's meeting has already concluded. It was a short one. But the meetings begin at 2pmET/11amPT, typically on Mondays. If you'd like to join us in the future, you would join the CircuitPython audio channel towards the bottom of the list on the left side of your Discord window (on desktop Discord anyway) during the meeting time.
idle owl
I would like to attend the meeting. This would be the first time. How can I join...
If you would like to be notified about future meetings, I can add you to the circuitpythonistas Discord role. We ping a couple of times per week about the meeting and once in a great while with other things.
But that is entirely optional and up to you.
crimson basin
cool
idle owl
@onyx hinge Do I really want the LED animation library to be dependent on another library? I think I'm still a little unclear on what this is solving. https://github.com/adafruit/Adafruit_CircuitPython_LED_Animation/pull/81
crimson basin
yes
crimson basin
thank you kattni
idle owl
You're quite welcome
onyx hinge
@idle owl I'm not sure if it should be accepted or not. The idea is, adafruit_ticks is on all boards and doesn't have accuracy issues on m0 boards after they've been running (float precision limitation) .. and it gets the nasty if/try/import/etc out of the LED animation library. That doesn't mean it's necessarily a good idea.
idle owl
<@!330227457296957440> I'm not sure if it should be accepted or not. The idea i...
Hmm. Fair enough.
onyx hinge
It was also a good exercise for adapting existing code to adafruit_ticks and putting it through its paces, so a learning experience for me
idle owl
It was also a good exercise for adapting existing code to adafruit_ticks and put...
Also excellent.
Here is the notes document for Mondayโs 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 youโll be attending the meeting - itโs super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and weโll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/1fsekWWLvcaMb7w6lfiDHqOjYbM38JgLOOxEjGdwh3A0/edit
storm minnow
Random question... How come the meetings are during business hours on a Monday?
stuck elbow
- they are only during business hours in your timezone, 2. many circuitpython developers do this as their day job
storm minnow
Fair and fair
stuck elbow
time is tricky
storm minnow
I guess "US business hours" would have been more accurate XD
idle owl
Random question... How come the meetings are during business hours on a Monday?
It started as a a sync up between paid CircuitPython developers, and eventually morphed into a community meeting.
stuck elbow
would have been easier if we didn't live on a spinning ball
idle owl
So initially it was a work-related thing.
storm minnow
Ahh...
idle owl
And was meant to be a sync-up for the week for us.
But folks started asking to listen in, and eventually we got others involved, etc.
storm minnow
shakes fist at job for taking that timeslot
idle owl
So it stayed at the same time/day, since we still need to sync up, but turned into a community effort as well.
blissful pollen
*shakes fist at job for taking that timeslot*
I'm more lucky in the fact my day job normally doesn't have meetings at the same time (knocks on wood)
idle owl
I figured there was a good reason for that seemingly random time XD
We're on opposite sides of the US, so it needed to be a reasonable time for each of us involved in the beginning.
storm minnow
I'm more lucky in the fact my day job normally doesn't have meetings at the same...
Sadly my job is being on the phones, so I don't really get any peace
stuck elbow
It's in the evening for me, but I rarely have any energy left to join it.
every time would have its downsides
storm minnow
Energy hard... a precious resource
idle owl
If any of the CircuitPythonLibrarians have a moment, I'd appreciate a review: https://github.com/adafruit/Adafruit_CircuitPython_Bundle/pull/332 Thanks!
idle owl
@onyx hinge This works either way, but we went with the .git address for the .gitmodules file in the Bundle. Please update it at your convenience. Nothing else needs to be fixed along with it. https://github.com/adafruit/Adafruit_CircuitPython_Bundle/blob/094622fbd46b719e264cab2632f60ab97ca12ccf/.gitmodules#L841
slender iron
@idle wharf I just copied the last one up
idle wharf
<@!624005998696333331> I just copied the last one up
Woot! I'll look at it later tonight and see what sort of markdown magic I can work up
storm minnow
So, kattni said that the "good first issue" tag isn't really well maintained for finding first issues... What's the best way to do so? Is there a way to see all pending reviews?
idle owl
So, kattni said that the "good first issue" tag isn't really well maintained for...
Not sure I follow what you're asking. Well, I followed it up to the pending reviews bit, anyway.
storm minnow
Not sure I follow what you're asking. Well, I followed it up to the pending revi...
When I'd asked about where to start, you said We're currently really bad about labeling "Good First Issue"s. So don't use that to pick something. We need to be better with that. So I wasn't sure how to best find reviews I could at least spellcheck or smaller issues I can grapple
idle owl
When I'd asked about where to start, you said `We're currently really bad about ...
Oh! You mean reviewing PRs. Those don't have the "Good First Issue" label, that's for issues only. In terms of reviewing, you simply have to go through the open PRs and take a look at them. In terms of issues, unfortunately, until we go through and curate better, the only way to find an issue to take on is go go through the issues and see what you speaks to you. Terrible answer, I know, but it's the state of things at the moment.
PRs aren't really labeled in terms of good first issue etc. If they're labeled at all, it's more to categorise the contribution than it is to assist reviewers.
storm minnow
Ah, yeah, I meant PRs. Still getting used to terminology XD I've realized that my development skills are on an entirely different plane of existence than the dev work y'all have been doing
idle owl
Ah, yeah, I meant PRs. Still getting used to terminology XD I've realized that m...
We all started somewhere! And I don't say that flippantly. Before July 2017, I had never touched programming or electronics.
idle owl
Ah, yeah, I meant PRs. Still getting used to terminology XD I've realized that m...
So for PRs, you can take a look at the list here: https://circuitpython.org/contributing and basically go through the links one by one and see if there's anything on the level you're looking for, in terms of spellchecking or syntax etc.
storm minnow
We all started somewhere! And I don't say that flippantly. Before July 2017, I h...
Oh, 100%! I mean, there was a time when I couldn't build a computer, and now I'm the one everyone asks for help in life XD
storm minnow
So for PRs, you can take a look at the list here: https://circuitpython.org/cont...
Ok, cool. GitHub's pages were a little confusing XD I ended up on some branch and had no idea where I was
idle owl
Ok, cool. GitHub's pages were a little confusing XD I ended up on some branch an...
Yeah GitHub isn't the best way to go about it, which is why we aggregated it. We had to. We have 280 some libraries, trying to keep track of them all without our Adabot helper would be impossible.
Those links will take you to GitHub. But that's the best way to get there.
storm minnow
Thanks ^_^
river quest
8-6-2021 is the Snakiest day of the year and itโs also this yearโs CircuitPython Day! The day highlights all things CircuitPython and Python on Hardware. Working with CircuitPython? Tag your projects #CircuitPythonDay2021 on social media and Adafruit will look to highlight them.
river quest
8-6-2021 is the Snakiest day of the year and itโs also this yearโs CircuitPython...
onyx hinge
@river quest posted & pinned in #help-with-circuitpython too
nice video, I never knew circuitpython was so easy to use
slender iron
manic glacierBOT
hollow token
Sorry I've been away from this for a while but I finally got around to https://github.com/adafruit/Adafruit_CircuitPython_MCP230xx/pull/41
onyx hinge
@tulip sleet an alternative: Find the top directory of circuitpython relative to the script -- https://gist.github.com/c63c05985bc41cd08cfe54e6f11d03a2
tulip sleet
Feel free to push that. I'm still wondering why it ever failed at all.
onyx hinge
@tulip sleet sure, I didn't want to interfere further with that PR so I made a fresh one; I'll rebase it after 5069 goes in
remembers it's amazing to have CI at all .. just think how it used to be
tulip sleet
i find pushing commits to someone else's PR to be a tricky business if it's not possible to do from github. The hint that git gives is not quite right.
onyx hinge
it's true, I'm also gunshy about that.
manic glacierBOT
analog bridge
The recent addition of BusDevice on Proxlight Trinkey M0 (PR#5058) doesn't leave enough room for the traceback module.
It is the only board that is failing, what should I disable on it BusDevice or traceback?
tulip sleet
The recent addition of `BusDevice` on `Proxlight Trinkey M0` (PR#5058) doesn't l...
BusDevice is more important on that board
we might find something else to remove later. I already added the tricky flag that turns off SPI
idle owl
@tulip sleet Am I ok to merge ^^? You tagged Jeff and Scott on it as well.
@tulip sleet Regarding https://github.com/adafruit/Adafruit_Learning_System_Guides/pull/1642 - foamyguy did it this way in the CLUE code that was updated: https://github.com/adafruit/Adafruit_CircuitPython_CLUE/pull/48/files (or rather it was already done that way and he simply updated it to poll the button press), but my point is, is the while button: bit needed in the Pyloton code? Or is the single button press check enough?
tulip sleet
<@!329766224093249548> Regarding https://github.com/adafruit/Adafruit_Learning_S...
I think the single check is good enough. It waits 0.1 seconds, which handles any bounces. Next time around the loop, if it's still pressed, the volume will be adjusted again. https://github.com/adafruit/Adafruit_CircuitPython_CLUE/pull/48/files replaces one while with another while. The Pyloton fix replaces an if with an if.
idle owl
I think the single check is good enough. It waits 0.1 seconds, which handles any...
Right on.
tulip sleet
<@!329766224093249548> Regarding https://github.com/adafruit/Adafruit_Learning_S...
Fine to merge. I would really like to have AND and OR for reviewers
thanks!
idle owl
Done and done.
idle owl
@lone axle You around for a max_glyphs question?
idle owl
@lone axle Ok, so I'm using the MacroPad library. Which uses Simple Display Text. I can't find max_glyphs anywhere but I'm getting Please update your code: 'max_glyphs' is not needed anymore. when I run my code. What am I missing?
Is it buried in Simple Display Text or something and I missed it?
lone axle
Is it buried in Simple Display Text or something and I missed it?
that would be my guess. Looks like the removal PR was merged: https://github.com/adafruit/Adafruit_CircuitPython_Simple_Text_Display/pull/4 but there hasn't been a release yet, so if you have the version from the bundle it probably still has max_glyphs in it.
idle owl
Oh!!
That's why I can't find it.
I was searching on GitHub.
Ok, I'll do a release. Thank you so much! I was so confused.
lone axle
I just did the same and was confused for a minute also ๐
idle owl
OK, release made. Whew. Thanks again.
Oi.... downloaded a Project Bundle for the MacroPad library, and it downloads WAY more libraries than I expected. I am confused.
Oh ... wait.
checks a thing.
Ok found two of them. Still not sure where another one is coming from.
Oh nevermind. Wasted time. It downloads all libraries necessary for all examples present in the repo. Which includes the LED Animation example that Tim included 10 days ago. It's not based on a requirement, it's based on an example.
lone axle
Ah, that is good to know. I didn't realize it included requirements for examples in addition to the library itself.
idle owl
Ah, that is good to know. I didn't realize it included requirements for examples...
Only when you're downloading a project bundle that is embedded into Learn out of a library repo. Or potentially if you have multiple files in the same directory in the Learn repo. That's why we've switched to using folders with code.py in them, so it only includes the files you need. But we don't do that in library examples in the library repo. And I don't think we intend to start. So the Project Bundle for those examples will sometimes include things you don't need.
idle owl
@onyx hinge This may be in your wheelhouse. Or not. https://forums.adafruit.com/viewtopic.php?f=60&t=181598
onyx hinge
nope, "OV2640 SPI" is a different thing than what I've been working on @idle owl
idle owl
Ah fair enough. I'm no help either.
onyx hinge
somebody makes this product "arducam" that puts a microcontroller between you and the camera; you can talk to the microcontroller via spi and ask it to talk to the camera. beyond that I don't know much about this product
https://github.com/adafruit/Adafruit_CircuitPython_OV2640/issues/10 they asked a similar question over here, which I responded to & closed
idle owl
Ah. Huh ok.
idle owl
@lone axle Good morning! Are you around for a question you may or may not have an answer to?
Nevermind, I fumbled my way through it.
idle owl
@tulip sleet So.... stumbled upon the fact that you apparently can't import usb_hid without it trying to find some kind of connection. Which means you can't power MacroPad projects using the MacroPad library from a battery or wall plug because it fails with USB busy (I think that's the message, I can only see what's on the display). Is there a way around this? Can I check for whatever CircuitPython is checking for, and not import it if it's not present... or I'm tossing things out here, I have no idea.
tulip sleet
i'll look at that, it should be innocuous just to import it
idle owl
I did a demo that does not use HID. And it fails to run on wall power.
I'm guessing about it being usb_hid. Might be something else.
idle owl
I need to give you the file, it's not on GitHub yet.
Yeah it could be..... but what?
I don't know how to troubleshoot it even.
Not very much serial console shows up on the display.
tulip sleet
Macropad() does self._keyboard = Keyboard(usb_hid.devices). it's probably failing there
tulip sleet
so check whether self._keyboard is None, and if it is, import that stuff?
idle owl
Except if you called keyboard and keyboard_layout you'd call _keyboard twice and it would fail again.
tulip sleet
you can do some kind of "is this the first time?" for .keyboard and .keyboard_layout,
tulip sleet
you could still import all the keyboard stuff, just don't create the Keyboard() until you need it. You don't have to conditionalize the keyboard imports, though it would save space to import them only on demand.
like def _touch() in the CP library, for instance (it does not do conditional imports, just existence checking)
idle owl
That's what I thought you meant.
I'm more fuzzy on exactly what it looks like. Committing what I have so I can git stash drop if I screw if all up.
Ok what I did at least lets the demo work.
Now to see if HID still works.
onyx hinge
On my calculators, I've made it so the use of USB HID is postponed until the paste function is first used, but the way I did it is not necessarily a good one. Let me find a link to the source..
onyx hinge
idle owl
Seems to work!
need to try keyboard layout
Works!
Now to fix MIDI since it has the same issue.
idle owl
@tulip sleet Can you review this? It includes the lib change. https://github.com/adafruit/Adafruit_CircuitPython_MacroPad/pull/13
idle owl
Thank you!
manic glacierBOT
tidal kiln
@idle owl hey - i think you know this one? sphinx issue when you have single quotes in a docstring for trying to do code markup, just add more quotes or something?
idle owl
<@!330227457296957440> hey - i think you know this one? sphinx issue when you ha...
Do you mean backticks?
tidal kiln
yep
idle owl
Yeah double on each side. Sphinx ignores.
tidal kiln
so change a to a ?
meh...forget how to escape markdown here...
`a` change to ``a``
^^ that?
ah, yah, nvm, that's it...found an example in another lib
@idle owl got it. thanks.
idle owl
Ah sorry. Yeah.
slender iron
cp as a UART bridge:
import board
import busio
print("reading uart")
serial = busio.UART(rx=board.RX, baudrate=115200)
while True:
print(serial.read(serial.in_waiting).decode("ascii"), end="")
I was doing two things at once and didn't realise you were asking more questions. Glad you got it sorted!
idle owl
the error message for that is super non-obvious
Yeah it's super obtuse. It tries to import or run the thing in single backticks, so you think it's an automock issue or whatever, or why is it failing on a docstring, etc.... Entirely unclear.
idle owl
@slender iron If you haven't already, are you going to share the multi-status-pixel love with NeoTrinkey as well? I might have a guide or two to update with these changes. Not going to do that until 7 is stable though.
jaunty juniper
too bad the RGB LEDs are on I2C on the keybow ๐ฆ
slender iron
@idle owl done
idle owl
<@!330227457296957440> done
Thanks!
idle owl
@onyx hinge What became of the auto-generated CIRCUITPY-contents screenshots?
onyx hinge
@idle owl They exist
idle owl
How existential.
onyx hinge
they are supposed to be updating daily, but I haven't checked lately if it's happening as planned.
idle owl
Hmm. How do I know what I'm looking for though?
Oh I kind of see the naming scheme.
Can I use them?
idle owl
Limor said they're "on pause". So I ended up downloading the image and uploading it directly from my computer for now. She wasn't clear on whether I should or should not use them...
onyx hinge
where is the place that this should have been documented?
idle owl
I have no idea.
onyx hinge
really? hm
idle owl
Wait
I do have an idea
in the ALS Academy guide
I mean current authors aren't going to look there intuitively
but we can post to Slack and email the team and let Anne know, and then everyone has a place to look.
idle owl
really? hm
If you know something she doesn't, please mention it on Slack.
I'd like to be using them, because if I had been, this py/pixelbuf update wouldn't have been necessary.
Would have been dynamic.
onyx hinge
right
idle owl
@onyx hinge Oh hmm. It's designed to work with Bundlefly specifically then? So to generate the image, it has to be a directory/code.py situation, not a bunch of files in the same directory - that only generates a single image for all of them?
I mean that's what I'm seeing, but I'm verifying with you.
onyx hinge
yeah I think that is correct.
idle owl
Ok I have to update the template code then.
onyx hinge
Does it match what the zip contains?
idle owl
There's no image for some of them at all. But they're a bunch of files in a single directory.
Some of them are near-pseudocode example though that you're supposed to adapt.
So I had them loose in the directory.
But I need images.
onyx hinge
I'd have to review what it does, do you want to give me a specific example?
idle owl
Um.... I think it's a me problem. I don't think we want to change how it works to deal with templates only.
When I should be able to adapt the template examples pretty easily to cause images to generate (in theory)
If not, it's not a huge deal, at least templates are a single point of failure for updates, versus a bunch of separate pages.
onyx hinge
OK, if there's a specific .py file you thought it should have made an image for, and yet it didn't -- let me know
or where you think the content of the image is wrong
idle owl
Seriously itching to update this template code to use rainbowio, but I would also have to put a warning in the template about using CircuitPython 7. I guess I probably shouldn't do that. Or I can put mangley try/except import blocks in. But it has to account for 6x, boards without _pixelbuf, and rainbowio. So.... super mangle. Bleh.
tulip sleet
Seriously itching to update this template code to use rainbowio, but I would als...
If it's template code, you could put in separate 6 and 7 versions of the code, and then later remove the 6.
since template code is just copy/pasted (if I understand this)
idle owl
If it's template code, you could put in separate 6 and 7 versions of the code, a...
Eh, the templates are designed to be "tailored". And doing that kind of defeats that purpose.
idle owl
since template code is just copy/pasted (if I understand this)
Not exactly.
slender iron
@tulip sleet what was the buffer issue you found?
idle owl
The reason I want to update is right now it's a file on the Learning Guides repo where you have to update the PIXELBUF_VERSION in the code to match what the board has. If I used rainbowio I could embed it in the template and be done with it.
As it is now, every guide has its own copy of the code. It's annoying.
idle owl
Too bad we can't backport rainbowio to 6.
tidal kiln
thinking about removing the big red warning boxes from the displayio guide that say to use latest in the example section:
https://learn.adafruit.com/circuitpython-display-support-using-displayio/examples
thoughts?
idle owl
thinking about removing the big red warning boxes from the displayio guide that ...
Do it.
There are probably warnings I should also remove in places. But that's a whole other ordeal to find all of those. I'm in template land at the moment.
tidal kiln
thanks. i'm in there now making some other updates from feedback. and noticed those. guessing very few people are using anything before 4.x these days.
idle owl
thanks. i'm in there now making some other updates from feedback. and noticed th...
One hopes.
tulip sleet
<@!329766224093249548> what was the buffer issue you found?
https://github.com/dhalbert/circuitpython/compare/main...debug-audio#diff-7f2729ae33d8da45b91ced94cc11b8e78402c08a6bec0b894fb0ad10f2328a5c starting around line 235, The m_realloc() for the first buffer was never called if no conversion was necessary, so it first_buffer was NULL the first time, it reused the buffer from the WavFile reader as first_buffer, due to code at line 141 in that file.
It's not completely clear to me this is definitely causing problems, but it could cause as use-after-free issue, because the file reader would be unaware someone else was hanging on to that buffer.
It was suspicious because the code for first_buffer and second_buffer was not the same.
Also I just fixed a bug that 8-bit audio had half the samples removed, so, e.g. a 1-second sample played only for 0.5 second (buffer size was not doubled when converted)
output is still scratchy at higher sample rates when display is in use. Also the USB connection goes away randomly.
I saw that before. It is not a crash, just USB going away
the new code has some belt-and-suspenders checking in it to make sure we're not writing past the end of some buffers. I may remove that later.
I don't see signs yet that buffer filing is getting behind. I'm checking for both channels being unbusy and that only happens at the end.
slender iron
๐
sounds like good progress
and a failure in me trying to prematurely optimize
tulip sleet
also I did some checking to make sure the data in the buffers is not garbage, and it has the correct values (with my non-dithered square waves)
there's still a lot of "why is it doing that???"
still don't understand the scratchiness. I did Saleae the I2S output, but it's hard to check, and the Saleae parser is seeing what it thinks are errors due to clocking glitches (even if I deglitch the recording).
slender iron
hrm
tulip sleet
i will try Saleaeing the PWM output with clean square waves, which should alternate between two duty cycles
slender iron
right
analog bridge
@slender iron where would the implementation of empty traceback object type go? py/objtraceback?
idle owl
Hmm. I just realised the way I updated the pixelbuf implementation means that the bundler isn't downloading pypixelbuf anymore anyway. Meaning for the M0 non-express boards, there's no colorwheel available to 6.x CircuitPython anymore if folks don't specifically grab pypixelbuf from the bundle. NeoPixel and DotStar still work with pypixelbuf, but the Bundler isn't grabbing it anymore.
tulip sleet
hmm, not sure how to fix that without adding some version-specific functionality to something
manic glacierBOT
idle owl
hmm, not sure how to fix that without adding some version-specific functionality...
I don't think it's worth fixing. I will update the four references to it in Learn to use rainbowio and update the templates to use rainbowio as well. I was going to update them to a try/except import block but realised there's no point for the M0 boards. There aren't any templates in M0 non-express guides yet anyway (that don't have _pixelbuf built in anyway....) I guess I could try except rainbowio and pixelbuf, leave off the pypixelbuf part.
@onyx hinge This is one of the instances: CircuitPython_Heart_Sculpture/code.py:3:import adafruit_pypixelbuf CircuitPython_Heart_Sculpture/code.py:16: pixel[0] = adafruit_pypixelbuf.colorwheel(hue)
I'll fix it, simply letting you know.
onyx hinge
idle owl
Yeah that's probably bad.
onyx hinge
that's the image that gets generated
idle owl
To have both.
@onyx hinge There's a guide for it right?
I'll add a 7.x warning to it.
Actually having both doesn't matter, but it never gets to pypixelbuf in neopixel imports. I guess you could still import colorwheel from it separately.
Maybe this is fine.
onyx hinge
dotstar imports pixelbuf now?
onyx hinge
and of course I am using pypixelbuf directly, for colorwheel pixel[0] = adafruit_pypixelbuf.colorwheel(hue)
onyx hinge
I think .. it's as good as it can be?
idle owl
To leave it for now.
Yeah.
What I want to know is will the bundler grab pypixelbuf if the import looks like this: py try: from rainbowio import colorwheel except ImportError: try: from _pixelbuf import colorwheel except ImportError: from adafruit_pypixelbuf import colorwheel
Only one way to find out I guess.
It's ugly but I'm done making the same example over and over for each guide when it could be a single example in the template.
onyx hinge
I don't know what it does in that case, and I'm also not sure whether it'll match what the image-maker gets since they are different bits of code
I do know how to run the image-builder locally so we could see..
idle owl
I know we've done it elsewhere, let me find some place we do it, and check the associated image.
oh
That could also work.
onyx hinge
it's better if you can check a zip file though
onyx hinge
rainbowio is a built in module, though?
or is there a bundle / pure python version too?
idle owl
hmm.
no
It's built in.
But I need something that excepts to a library
to see if the bundler grabs it
This is a really dumb way to do this.
I don't think I can test it.
Shrug.
idle owl
I got this image with that code
Is this related in any way to Bundlefly, or is it a completely separate mechanism?
onyx hinge
frustratingly, it's separate, so we don't know if they behave the same
idle owl
Hmm ok
onyx hinge
I assume that the bundlefly zip-maker is in the learn system somewhere, which is not on public github
idle owl
Right.
onyx hinge
I hope that at some level it's using similar code but we can't know for sure. it would be best if one program made the images and the zips, because then it would at least be consistent
idle owl
Asking Justin instead of speculating and wasting time searching for an example in the monumental beast that is the Learning Guides repo.
@onyx hinge Turns out bundlefly uses findimports.
Or rather, "that's how we start building the dependency tree, the rest is via the json that is generated that defines the imports for each lib"
CircuitPython_Templates/status_led_one_neopixel_rainbow/code.py: could not find rainbowio.colorwheel
CircuitPython_Templates/status_led_one_neopixel_rainbow/code.py: could not find _pixelbuf.colorwheel
code:
_pixelbuf.colorwheel
adafruit_pypixelbuf
board
neopixel
rainbowio.colorwheel
time```So, yes, it will work.
Perfect.
onyx hinge
findimports is what the image maker uses too
idle owl
@onyx hinge Excellent. Then there is much reason to believe they will behave the same.
idle owl
@lone axle Is there a way to dynamically change the content labels to swap from white text on black to black text on white using https://github.com/adafruit/Adafruit_CircuitPython_DisplayIO_Layout/blob/main/adafruit_displayio_layout/layouts/grid_layout.py ?
storm minnow
Just flashed CP 7 alpha 5 to MacroPad, it's going to a "raw REPL", can't connect to serial with PuTTY and get "access is denied" -- did I find a bug?
idle owl
Updating mine, hold on.
Or maybe it was already up to date.
Because I already downloaded alpha 5 previously.
And my code is still running.
idle owl
Just flashed CP 7 alpha 5 to MacroPad, it's going to a "raw REPL", can't connect...
Not sure what's going on there.
storm minnow
Hmm... Anything I can try?
idle owl
So you start up and it automatically goes to the REPL?
storm minnow
After running the inbuilt Hello World demo
idle owl
Anything on it you need to keep? I would try the nuke.uf2 on the CircuitPython page in the guide first, try flashing it, and then load CircuitPython again and see if it's still doing it. I mean, that's a burn it to the ground and start over plan, but it's not necessarily a bad one.
@storm minnow This: https://learn.adafruit.com/adafruit-macropad-rp2040/circuitpython#flash-resetting-uf2-3091244-20
lone axle
<@!382939733107408897> Is there a way to dynamically change the content labels t...
I think you can access the objects inside the GridLayout using square brackets with index so something like
my_grid_layout[0].text_color = 0x000000
my_grid_layout[0].background_color = 0xFFFFFF
Would do it I think, though I've not tried it specifically.
idle owl
I think you can access the objects inside the GridLayout using square brackets w...
Oooh. Ok.
lone axle
index would be the order they were added I believe.
idle owl
Hmm ok
storm minnow
Anything on it you need to keep? I would try the nuke.uf2 on the CircuitPython p...
I had just flashed CP for the first time, so no problem nuking. Just figured I'd share the oddity
lone axle
a future enhancement might be to provide a function that allows accessing them via their x,y cell location within the grid though. That seems like it'd be more easily usable than index.
idle owl
I had *just* flashed CP for the first time, so no problem nuking. Just figured I...
Yeah for sure! Thanks. Try nuking it first though and see if it's a corruption issue.
If not, perhaps it's a bug.
idle owl
a future enhancement might be to provide a function that allows accessing them v...
Yeah maybe. Otherwise you have to wrap your brain around the order in which you created them, which in my case is done in a for loop, so it's not all listed out individually.
storm minnow
Here's video of exactly what happens:
storm minnow
Yeah for sure! Thanks. Try nuking it first though and see if it's a corruption i...
No problem :D oh, I had gotten into REPL once, but only once. Otherwise the 'raw REPL'
idle owl
Hmph. I'm stumped, but that doesn't mean others wouldn't know what's going on better.
storm minnow
If there's no like, logs I can obtain from the board, guess I'll just nuke and see if it happens again
idle owl
Not really yeah.
idle owl
I think you can access the objects inside the GridLayout using square brackets w...
text_color doesn't seem to be changing. The background is though!
storm minnow
Ahh, I timed connecting to it IMMEDIATELY after resetting [while code was running] and it went to the normal "code done running"
Hmmmm.... reset again and no more raw REPL... very strange
lone axle
`text_color` doesn't seem to be changing. The background is though!
ah whoops, I had the property name wrong. It should be just color instead of text_color
idle owl
ah whoops, I had the property name wrong. It should be just `color` instead of `...
Yep I checked something else and saw that. Got it!
storm minnow
Well, I gotta go, but I'll nuke later and let you know if it happens!
idle owl
Well, I gotta go, but I'll nuke later and let you know if it happens!
Thanks!
lone axle
I'll make an enhancement issue on GridLayout for the x/y cell location. That would be great functionality I think.
idle owl
I'll make an enhancement issue on GridLayout for the x/y cell location. That wou...
Brilliant. I used it in the Braille Keycaps code instead of the code PhilB wrote that generates all the shortcut labels and does a bunch of math to spread them out. So I'm trying to adapt it to having the MacroPad rotated 90 degrees. But displaying 4x3 shortcuts grid on the display is difficult in that orientation. So instead, I'm doing 2x6, and highlighting the shortcut when the key is pressed.
Ideally it would change the entire cell background. But it only changes the text label itself.
The way I generate the labels makes this work properly: py if key_event: if key_event.pressed: layout[key_event.key_number].color = 0x000000 layout[key_event.key_number].background_color = 0xFFFFFF
lone axle
Yeah, I think for that we need a different type of Label object inside display_text. The current ones are sized entirely dynamically based on the text that they contain. We need one that allows itself to be resized to arbitrary sizes rather than basing it off of the contained text. I started working on one like that at one point but didn't end up wrapping it up enough to publish.
idle owl
Yeah, I think for that we need a different type of Label object inside display_t...
Ah fair enough.
lone axle
That way the grid could control the size of the label to make it fill the cells.
idle owl
Nice
I mean this still works pretty well.
I'm pretty happy with it. Not sure Limor will want a guide for it because it doesn't match the key layout exactly. But if I can come up with something, I'll use it myself.
@lone axle ^^
Each set of 4 is a row.
Doesn't actually do HID right now because I'm terrible about pressing things randomly if they light up on my desk, so I try to avoid leaving HID running if I'm not explicitly using it ๐
lone axle
Nice! I had never considered using it at 90 degrees before but I do really like the way it looks now that I've seen it.
idle owl
Right?! Display is easier to see, knob's there if you need it.
I like 90 best of all the rotations I think.
MacroPad library maps the keys and pixels, and rotates the display.
The pixelmapping was the worst part of that bit. What a beast.
onyx hinge
@idle owl I wonder if you can fit dotted lines somehow to emphasize which ones are the groups of 4 .. is it the first 2 display rows are one physical row? (in any case, neat, it's nice that you can rotate the display during a program!)
idle owl
<@!330227457296957440> I wonder if you can fit dotted lines somehow to emphasize...
I was wondering the same thing about delineating the groups.
onyx hinge
do two of the groups in reverse video, then choose keycaps that are light/dark themed?
idle owl
Top line is a label. The rest are grid layout.
onyx hinge
you know more about that stuff than me
idle owl
Not sure what to use to generate a line.
@lone axle Could you also make it so you can add lines between the cells? And the ability to specify only some cells having an outline?
Or is that bonkers ๐
lone axle
display_shapes library or maybe vectorio can make lines I think. It would require a bit of manual placement to have them seperate the cells atm I think. I can look into providing that as an option within GridLayout to make it easier though.
idle owl
`display_shapes` library or maybe `vectorio` can make lines I think. It would re...
That would be amazing. I don't mind manually placing things for now. I'll take a look at vectorio. Thanks!
Seems to only offer shapes, not lines. Hmm.
Maybe a rectangle with no height?
I have literally no idea how to use this.
lol.
TypeError: unsupported pixel_shader type```Bleh, code runs, but nothing happens.
Have to run. Later!
hollow token
Thanks for all the help getting the SPI version of the MC23XXX chips in! :)
lone axle
Rectangle with 1 height for a line maybe, or maybe a polygon with just 2 points. I do think it would be great if there were some more example code for vectorio. Perhaps there is some out there that I am unaware of.
Maybe it would be good to add a section that covers it in the main displayio learn guide.
idle owl
Rectangle with 1 height for a line maybe, or maybe a polygon with just 2 points....
I can't figure out how to use it. There's no show() or anything.
I see the little Blinka, so it's not displaying anything.
lone axle
IIRC you create your object either Rectangle, Circle, or Polygon. Then you add that object to a VectorShape object: https://circuitpython.readthedocs.io/en/6.3.x/shared-bindings/vectorio/index.html?highlight=vectorio#vectorio.VectorShape and then that VectorShape can be added to a Group in order to be shown.
but it's been a little while. I wrote some sample code at one point that is on one of my devices and I always have to refer back to it because I never managed to get it committed to memory yet.
idle owl
Oh!
Add to a group.
Ok
I now have a blank spot, lol. So it drew something alright.
I don't get how to make it show up white.
Anyway I need to bail for the evening. Thanks for your help!
jaunty juniper
Just flashed CP 7 alpha 5 to MacroPad, it's going to a "raw REPL", can't connect...
that looks a lot like you have some other software running on your computer doing something, are you using Cura for example ?
storm minnow
that looks a lot like you have some other software running on your computer doin...
I do have Cura running in background, but it doesn't connect to anything on serial
jaunty juniper
it sure does, unless they fixed it
what if you quit it ?
though it usually manifests as M105 being sent to the REPL, not the raw REPL, it also manifests as the serial port not being available
storm minnow
I still have Cura running, but it isn't doing that raw REPL thing now
What is 'raw REPL' anyway?
jaunty juniper
it's like the REPL but for computers, tools like thonny use it
storm minnow
Ohhhh, so for software to have serial comms with CP devices? noice
jaunty juniper
it doesn't interpret what you type (like tabs as completion) and doesn't echo/output anything
it's ctrl-A, paste mode for humans is ctrl-E
storm minnow
Wonder what was kicking it into that before... Unless it was Cura, and it's just intermittent
jaunty juniper
some software hogging the serial port no doubt, did you look at the link above ?
(solution at the bottom is to disable USB printing in cura)
without any input from me it goes into the REPL after approximately 30 seconds. , and I see the following (transcribed, may be some minor typos) "Adafruit CircuitPython 4.0.0-beta.4-1-27-g67440acd3 on 2020-02-03; Adafruit CLUE nRF52840 Express with nRF52840 >>> raw REPL; CTRL-B to exit >
storm minnow
Hadn't gotten to the bottom yet XD I have a feeling it's not Cura, as I've got it sitting in REPL now and nothing is happening... And Cura isn't configured to print to USB at all
I wonder if it might be the pump...
jaunty juniper
I don't know, same exact issue and you happen to be running cura, I know where I put my 2 cents
storm minnow
Hmm, their pictures show regular REPL...
jaunty juniper
yeah I suspect on windows the use of the serial port is more exclusive, so once putty is running cura can't access it, on my mac I would see it output M105 while I'm in the REPL, might be a difference in the way screen or tio access the serial
it doesn't need to be configured too, it just hogs the serial ports, out of the box
storm minnow
Hmm
It stopped doing it randomly, even after PuTTY was closed/disconnected... And I haven't seen M105 at all
jaunty juniper
It stopped doing it randomly, even after PuTTY was closed/disconnected... And I ...
I love that bug repport: https://github.com/Ultimaker/Cura/issues/5207#issuecomment-741814775
storm minnow
I love that bug repport: <https://github.com/Ultimaker/Cura/issues/5207#issuecom...
๐คฃ Oh gosh that's beautiful
storm minnow
@jaunty juniper I'm disabling USB printing since I don't use it; should I ping you if it happens again later?
jaunty juniper
not me particularly, but whatever it is, I think it's most likely something using the serial port if putty says "access denied" (as opposed to not found or timed out), so look at what is running when it happens
storm minnow
Okie dokie Loki!
analog bridge
Print failure info was removed from Build CI in #4720... was this change intentional? https://github.com/adafruit/circuitpython/commit/446169e07869386cef9658d594bb3fbde56e94c8#diff-5c3fa597431eda03ac3339ae6bf7f05e1a50d6fc7333679ec38e21b337cb6721
onyx hinge
`Print failure info` was removed from `Build CI` in #4720... was this change int...
Not likely? Open an issue?
fluid canopy
Hello, thank you very much for the invitation, I have a doubt I want the example that you created in python Adafruit_CircuitPython_RockBlock / examples / rockblock_send_text.py can be displayed in a window, I don't know anything about python, only the most basic thing, I would appreciate it a lot if it can be done
idle owl
@tulip sleet Do you know who would know how vectorio works or is supposed to work? Because it is not working the way it feels like it should be, and I have no idea where to even begin.
tulip sleet
WarrierOfWire started on it, and kmatch98 has also worked on it. I'm assuming you read the guide on it.
idle owl
There's a guide? Ugh. I missed that.
I checked Learn....
Yeah I did check Learn. I didn't find anything.
tulip sleet
sorry, I was mistaken about that -- I thought there was a guide (there should be, or more pages in the displayio guide).
idle owl
Yeah that might be a good place for it.
But I'm not the one to do it apparently. Since I can't even figure it out.
tulip sleet
there are a number of people who've used it in #help-with-circuitpython , and i've seen queries, so i think going there with a use case will help.
idle owl
Alright, thanks.
tulip sleet
i haven't used it at all, sorry
idle owl
No worries. I didn't think you had, but I thought you might know who had. You did ๐
uneven mango
I use circuitpython.org filters to find boards for projects. I'm guessing others may too. I'm thinking of creating a PR for circuitpython.org to add a filter for a breadboard friendly feature. Defining bread-board friendly as two parallel pin rows with at least one breadboard socket available on both sides of the board. Horizontal pins could include UART/SWD, but not critical pins like power, ground, GPIO, etc. This would allow filtering out boards like the pyportal, magtag, funhouse, etc. Before I start researching all the current boards, is this a worth while idea?
idle owl
I use circuitpython.org filters to find boards for projects. I'm guessing other...
@gilded cradle What do you think? ^^
idle owl
I use circuitpython.org filters to find boards for projects. I'm guessing other...
It's a valid idea, I simply want to make sure we're not bloating the filter list to an unusable point. That's why I asked Melissa to weigh in. Please be patient and await her consideration. Thanks!
uneven mango
It's a valid idea, I simply want to make sure we're not bloating the filter list...
No problem. Completely understand the bloat factor.
tulip sleet
No problem. Completely understand the bloat factor.
Arduino-shape (e.g. Metro boards) is not breadboard friendly, but it's kind of its own breadboard. I wonder how you would characterize that. It's sort of purpose built vs general-purpose, but that's not very informative
same for Pi-shaped boards - not breadboard-friendly without an accessory, but "general purpose")
uneven mango
Arduino-shape (e.g. Metro boards) is not breadboard friendly, but it's kind of i...
I was thinking boards that can plug into a breadboard for prototyping. I see your point on some boards being like a breadboard.
ember iris
Arduino-shape (e.g. Metro boards) is not breadboard friendly, but it's kind of i...
I'd guess the easiest way is a form factor class?
spare prawn
i saw a mention of "readinto" for ble no longer taking the nbytes parameter for CP 4. All later documentation seems to support nbytes. I am getting a "ValueError: length argument not allowed for this type" in nordic.py line 77 when i try to use "readinto" with ble...with buffer as a memoryview and as a byte array, with and without nbytes set implicity or explicitly. same code works for other streams (busio.uart, stdin).
tulip sleet
there are "development" boards, meant for general-purpose use. That is an Adafruit shop category. But it includes CPX and other non-breadboardy boards
spiral elk
You'd also need to factor in distance between rows. There are some ESP8266 boards that are only useful by straddling across 2 breadboards.
tulip sleet
so that may not be such a useful distinction
uneven mango
You'd also need to factor in distance between rows. There are some ESP8266 board...
I considered that when defining breadboard-friendly as one socket available on each side, but assumed single breadboard.
gilded cradle
@idle owl that sounds useful..
idle owl
@uneven mango Figure out the best way to categorise things, and go ahead and make a PR.
spiral elk
And there also exist breadboards with a 0.6" center gap. radio shack sold them.
tulip sleet
so "breadboard-friendly" or "breadboard-pluggable", also maybe "general-purpose" (for CPX), "Arduino form factor" (not sure about exactly what to call that)
gilded cradle
Breadboard-Friendly sounds good
uneven mango
I think people would want to search for breadboard-friendly and arduino form factor. I don't know if "general purpose" for the board not meeting the other criteria would be searched for.
I'm thinking of mapping a project requirement to a board feature.
tulip sleet
i saw a mention of "readinto" for ble no longer taking the nbytes parameter for ...
could you show the bad line with the error?
gilded cradle
How about Arduino Shield Compatible?
Cause Arduinos are now all sorts of form factors
spare prawn
could you show the bad line with the error?
sorry i deleted message here beccause i reposted on the CP dev channel. here is the snippet
byte_count = device.readinto(memoryview(in_buf)[start_ptr:])
tulip sleet
that is a slice, not sure if it's supported in general in that kind of context (does slice have buffer protocol)
what kind of object is device
is this in CircuitPython or bleio blinka?
spiral elk
Question while people are awake: If I were to want to enable support for a board variant (same PCB, just a uC with a different amount of ram), would it be best approached as a compile-time option or a new board definition?
spare prawn
that is a slice, not sure if it's supported in general in that kind of context (...
device in the error case is UARTService()
spare prawn
that is a slice, not sure if it's supported in general in that kind of context (...
i also tried with a "bytearray" which i thought was a valid buffer type. same error.
stuck elbow
@spiral elk do you actually need a board definition for this? different LD file?
spiral elk
just a different ld file mainly
tulip sleet
which chip?
tulip sleet
that is a different board, since the CPU is different (and especially since STM differences are quite idiosyncratic). If this is a one-off, then you can keep your own fork for this. If it's a product, then we can consider a board defn.
spare prawn
that is a slice, not sure if it's supported in general in that kind of context (...
CP on Nordic Itsybitsy
tulip sleet
I am not sure if we want to have board defns for the huge array of cheap-o BlackPill clone variants or not
spiral elk
Well, this board isn't a clone, its from the original manufacturer
tulip sleet
the chips are enough different (from a search): _
The STM32F411 is backward compatible with the STM32F401 and should run your code for the STM32F401 without any modification.
The STM32F411 has a few additional features, so if you use them, your code will not work anymore on the STM32F401.some diferences:
new peripheral (SPI5) higher maximum frequency (100MHz instead of 84MHz) and differend voltage scales configuration larger SRAM Additional peripheral pin mapping for flexibility and for new peripherals Additional DMA remap for flexibility and for new peripherals
tulip sleet
i also tried with a "bytearray" which i thought was a valid buffer type. same e...
looking...
@spiral elk so definitely a new board definition
spiral elk
I think most of that is abstracted away in the STM HAL isn't it if MCU_<mumble> is conditionally set?
ionic elk
@tulip sleet yes boards are largely defined on the MCU they hold so the F401 blackpill should get a new def
spare prawn
looking...
thx!
tulip sleet
@spare prawn I think it is just a bug in Adafruit_CircuitPytyhon_BLE. It should no longer take a length. Could you open an issue? Does plain read() work for you in this case?
ionic elk
The blackpill also has "versions" where you add external flash but we didn't opt to make different profiles for that
@spiral elk just copy the og blackpill profile pretty much wholesale and swap out for the LD file we already have for the F401
and change the MCU flags ofc
spiral elk
Yeah, I have already done that and got it pretty much up and running already.
spare prawn
<@!680268218266681410> I think it is just a bug in Adafruit_CircuitPytyhon_BLE. ...
plain read worked before in other code..i have been refactoring to generaliize and trying memory view to improve performance...what i have works for other serial devices, just ran into this issue. i will now work to refactor for this case or just revert i guess.
ionic elk
@spiral elk nice, if you PR it I'll approve quick
spiral elk
Needs some cleanup, not sure if I can get to it this week.
tulip sleet
plain read worked before in other code..i have been refactoring to generaliize a...
honestly, I don't remember why it was removed
the issue will make us re-examine that decision
spare prawn
honestly, I don't remember why it was removed
can you point me to where the error message is generated? i tried to find it past nordic.py and could not find it.
tulip sleet
can you point me to where the error message is generated? i tried to find it pas...
I think it might be in _bleio.CharacteristicBuffer, but the complete backtrace should tell you
@spare prawn Please include the complete backtrace in your issue and also more context for the code, thanks. You can edit the original post
spare prawn
I think it might be in `_bleio.CharacteristicBuffer`, but the complete backtrac...
what do you mean by "complete" backtrace? is this suffiicient:
File "lib/metre/gAsynco.py", line 394, in _wrapper File "dev.py", line 234, in serial_task File "adafruit_ble/services/nordic.py", line 77, in readinto ValueError: length argument not allowed for this type
tulip sleet
@spare prawn if that's in, then that's it. I thought it would go deeper, but they way it works is tricky
onyx hinge
what do you mean by "complete" backtrace? is this suffiicient:
` File "lib/met...
@tulip sleet ``` def readinto(self, buf, nbytes=None):
"""
Read bytes into the buf. If nbytes is specified then read at most
that many bytes. Otherwise, read at most len(buf) bytes.
:return: number of bytes read and stored into ``buf``
:rtype: int or None (on a non-blocking error)
"""
return self._rx.readinto(buf, nbytes)
// CPython extension: if 2nd arg is provided, that's max len to read,
// instead of full buffer. Similar to
// https://docs.python.org/3/library/socket.html#socket.socket.recv_into
mp_uint_t len = bufinfo.len;
if (n_args > 2) {
if (mp_get_stream(args[0])->pyserial_readinto_compatibility) {
mp_raise_ValueError(MP_ERROR_TEXT("length argument not allowed for this type"));
}
len = mp_obj_get_int(args[2]);
if (len > bufinfo.len) {
len = bufinfo.len;
}
}
It seems that either that `readinto` needs to avoid passing `nbytes=None`, or the core code should treat `nbytes=None` like it was unspecified. I don't know which is more python-compatible. (`if (n_args > 2 && args[1] != mp_const_none` or so)i think this is due to CharacteristicBuffer or something like that. I don't remember why it was removed.
spare prawn
<@680268218266681410> if that's in, then that's it. I thought it would go deeper...
thanks for that pointer. I poked around and it looks like it might be a bug...i put some clues into the Issue.
spare prawn
thanks for that pointer. I poked around and it looks like it might be a bug...i ...
oh, i just saw @onyx hinge did the same thing...sorry missed that.
spare prawn
<@329766224093249548> ``` def readinto(self, buf, nbytes=None):
"""
...
I just tested changing the line 77 in nordic.py to return self._rx.readinto(buf). and it appears to have fixed the issue.
analog bridge
Not likely? Open an issue?
should I just go ahead and add it back?
onyx hinge
should I just go ahead and add it back?
@slender iron unless you had some reason for removing this, I think we should put it back?
analog bridge
`Print failure info` was removed from `Build CI` in #4720... was this change int...
Its regarding removal of Print Exception Info from CI. ๐
I was looking for it while trying to debug atexit test failure.
idle owl
@tulip sleet Can you get me a build of the audio fix PR? CI is still going to be ages and I don't work that much longer today.
For MacroPad.
ionic elk
@slender iron so the project that I'm worried about leaving dangling here is the whole "system reservation in sleep memory" thing, so that stuff like next_file can store information across deep sleep. I got decently far with it, (implemented deep sleep memory on STM32, created the API, had it compile and some basic tests working) but I've been bogged down in the debugging and haven't been in the right headspace to finish it with everything else going on.
slender iron
Its regarding removal of `Print Exception Info` from `CI`. ๐
I was looking for...
Ah, I might have copied micropython's change for it.
ionic elk
Is it ok if I put it down and come back to it informally later on? Should I post up what I have for someone else to pick it up? I could filter out the STM32 sleepmem stuff into a separate PR
slender iron
I think it's best to file an issue and then link to your WIP branch
that way it doesn't sit as a pending pr
ionic elk
Ok, that sounds good to me
slender iron
๐
tulip sleet
<@!329766224093249548> Can you get me a build of the audio fix PR? CI is still g...
This does not have the clock fix I gave to foamyguy, but it does have the audio fixes
idle owl
@tulip sleet Thanks.
idle owl
This does not have the clock fix I gave to foamyguy, but it does have the audio ...
Wav files are still super crunchy ๐ฆ
I used play_file from the MacroPad library.
tulip sleet
what is the sample rate?