@simple pulsar @slender iron mentioned that to me. My personal setup doesn't have an analog filter, I can't speak for what will be on the circuit playground ble
#circuitpython-dev
onyx hinge
raven canopy
audiosource.RawSample and audiosource.WavFile? Mixer doesn't quite fit there, does it? 😄
onyx hinge
@cerulean pawn I do know the speeds can be off a little bit, I prepared a small table that showed the frequency errors I expected . https://github.com/adafruit/circuitpython/pull/2000/files#diff-91d90d4c5b3b63bdcbfca203d293a5d0R48 claims that a 22025Hz sample would end up being played at 22038.5Hz (assuming the accuracy of the 16MHz underlying clock of the PWM)
simple pulsar
The Atmegatron uses analogue circuitry in the form of a steep 3rd order Chebyshev filter to maintain the high end, while still filtering off the modulation frequency. https://soulsbysynths.com/atmegatron/
raven canopy
also, this would be "fragmenting" by not having a common home, but couldn't WavFile be turned into wave: https://docs.python.org/3.4/library/wave.html#module-wave
slender iron
@raven canopy I like audiosource. I've been thinking about a synthio to handle envelopes and such too
raven canopy
hehe. conversely, i also like audiocore, because including Mixer makes more sense that way vice audiosource. (Mixer doesn't seem like a "true" source to me...since it has source children.)
ionic elk
@slender iron if you're picking up stm32s, just want to emphasize again how nice the DIS_F412 is! QSPI Flash, 1.5" touchscreen TFT, I2S audio codec, microphone, built in debug. We should be able to test most CPy features on it.
slender iron
@ionic elk I already ordered but added to my cart for next time
onyx hinge
I'm back if there's more to be said about audio.
swift arrow
on a grandcentral..I am using the example for the sd card..https://learn.adafruit.com/adafruit-grand-central/pinouts
I keep getting the following error..
OSError: timeout waiting for v2 card
thesd card is a 64 GB card formatted fat32
ohya.. using CP 4.1.0 RC1
doing the cardsneed to be formatted for 512 sectors?
meager fog
dev, dont forget its on its own SPI port
swift arrow
I;m using the sample code for it I thought.
argh...
the example should work right? or are you saying I need to set some special pins @meager fog
swift arrow
right that is what I am using
is 64GB just too big?
I used that because it is what I have laying around
tulip sleet
@slender iron I have reproduced https://github.com/adafruit/circuitpython/issues/2005 but only after several hours of running. Adding stacktrace info to the issue, it looks like a DMA issue. unfortunately we've lost a lot of the variables due to <optimized out>, but I got the contents of the DMA descriptors. If you have time to look and want any other values, I can retrieve them. I'll do another build with optimization turned off and run it overnight again.
swift arrow
I found a 32GB SDHC card andused SDformatter.. same issue
meager fog
dev, you could try it in arduino
slender iron
@tulip sleet I don't have the brainspace to context switch now
tulip sleet
np, I'll run it again with better debuggin on.
slender iron
kk
signal root
Are you using an open source library for the FAT in Circuit Python?
tulip sleet
@signal root it is this https://github.com/micropython/oofatfs
used by MicroPython and by CircuitPython
sand lotus
Oh wait, I guess I should have asked this in the help with section.
pastel panther
@prime flower, @graceful heart let me know when you guys chat about the WSGI lib
prime flower
@pastel panther np! @graceful heart would you be available sometime this next week to chat?
graceful heart
Yeah, maybe we could set a time some time to meet and chat about it?
pastel panther
sounds good to me
prime flower
What timezone are you in @graceful heart ?
graceful heart
East coast
pastel panther
I'm on the west cost but I'm generally available whenever
prime flower
@graceful heart Are you available to chat between 6PM and 7PM EST tomorrow? I could do later too.
graceful heart
Yes, tomorrow or today would work at those times
prime flower
@graceful heart Tomorrow works better for me, I have a meeting during that time today. Thanks for the PRs, they're incredible.
graceful heart
Great! @pastel panther does that time work for you? And of course, happy to contribute :)
stuck elbow
I just learned that someone made a game jame game for PewPew! https://www.epicgamejam.com/games/game-ego
ruby atlas
Resumed work on _pixelbuf to harmonize it's API with pypixelbuf. Naturally I have a crash on CPX that I don't have on the Metro M4. I guess I should grab a Metro M0 and a 10-pixel neopixel strip 😃
(which undoubtedly won't crash)
(ooh i think i fixed my crash!)
upbeat plover
FROZEN_MPY_DIRS += $(TOP)/frozen/feather_m4
i used this in my board.mk but it will not mpy all stuff in folders "adafruit_seesaw" and "adafruit_bus_device" only the "init.py" is being frozen
is there another way to use the "FROZEN_MPY_DIRS"?
upbeat plover
@slender iron did something change with frozen libs?
slender iron
@upbeat plover not that I know of
onyx hinge
@slender iron was the winning suggestion audiocore?
I have an existing clone which is giving me this error. I am at the tip of origin/master. Can someone tell me the recovery procedure? ```$ git submodule update --init --recursive .
error: no such remote ref 1dabc2df8ed1acaf0c483b58d757ba72e8b42090
Fetched in submodule path 'lib/tinyusb/hw/mcu/st/stm32lib', but it did not contain 1dabc2df8ed1acaf0c483b58d757ba72e8b42090. Direct fetching of that commit failed.
Failed to recurse into submodule path 'lib/tinyusb'
@onyx hinge core is fine with me. try git submodule sync too
onyx hinge
@dry vortex thanks, submodule sync was it. Will file away..
slender iron
great!
slender iron
it syncs the .gitmodules file to the .git/ copy
onyx hinge
OK, yay, the renaming wasn't as hard as I thought. Testing performed: build and check everything imports on a metro m4 express. Pushed, will PR it later this evening as its own thing.
manic glacierBOT
onyx hinge
hmph, failures on travis. sparkfun_lumidrive, pca10056, feather_nrf52840_express -- maybe all the nrf boards?
and feather_radiofruit_zigbee
and maybe that whole clan
well, I see the problem, don't care to stay up late enough to see the outcome on travis
tulip sleet
@upbeat plover the frozen module logic expects a repo at each directory. It may not work if the modules are too deeply nested. Try to imitate the directory structure of an existing board with frozen modules
signal root
What lib is used for the USB flash storage on M0/M4 chips?
tulip sleet
@signal root you asked the same question last night, I thought? Are you asking a different question?
signal root
Does oofatfs include the mass storage component?
Or just the file system component?
When I say mass storage the USB to SPI bridge?
tulip sleet
it's the filesystem part; the SPI or QSPI (or internal flash) part is implemented by our own code in CircuitPython. Are you looking for an Arduino solution? See this brand new library: https://github.com/adafruit/Adafruit_TinyUSB_Arduino which includes USB mass storage support
that uses the SDFat library underneath
signal root
Noice :)
I'm writing my own javascript platform, wanted to take cues from micro/ciruitpython
I flashed micropython on to the m4 saved a file and used arduino_spiflash to load it 😃
I'm binding arduino calls to in the JS
tulip sleet
see circuitpython/supervisor/shared for common code, and then see the supervisor code in each port/ we support for the device specific code
signal root
So that TinyUSB may be perfect
tulip sleet
ok! It's brand new, so let us know if you have issues
signal root
I mean this is pretty exciting if you can build your own platform with a couple of arduino libs
tulip sleet
yup! supporting USB storage on Arduino is a great thing
i've gotta sleep, but check later if you have questions.
signal root
Thanks @tulip sleet
tulip sleet
yw!
signal root
I'm using platformio and I don't think the platform has USB Stack enabled
signal root
got it building 😃
Got a filesystem saving to SPI! Boom! I am super shocked at how quick I've got everything up and running!
manic glacierBOT
stuck elbow
what does this mean?
exotic pumice
it means I got microchip to fix their stupid bad datasheet that destroyed me
this is what it used to look like
two 9s
main meteor
I remember some of the "which is reality" traffic, congrats! 🎊
exotic pumice
thanks
this is the perfect ending to my ridiculous story
I remember showing it to people and them not seeing it either, which is the scariest part lol
so glad it's fixed
I guess most people use atmel studio or arduino or whatever so it's not that big an issue
but it's egregious if you're writing your own hal
I wanna frame this darn datasheet
main meteor
Your efforts will save a bunch of people a bunch of head scratching, time and frustration.
main meteor
Bad datasheets are part of why I punted Microchip PIC for Atmel AVR back in the day (Microchip has improved since, which is good since they also absorbed Atmel).
exotic pumice
yeah I guess it was atmel that had the good ones
but they're atmel now so ¯_(ツ)_/¯
maybe the hashtag should say atmel but too late
or maybe I'm playing on the irony that america is about as great as microchip
who knows
main meteor
Back in the PIC days, there were two programmer protocols, one semi-documented, one proprietary. I couldn't afford the Microchip programmer, so I bought the Warp13 one instead. I didn't use DOS, so I used the open source command line tools. Then new chips came out and the Warp13 didn't have enough memory to support both protocols, so it then only supported the undocumented proprietary one 😠 so I could no longer use it if I wanted to work with the newer chips. That's when I surveyed the microcontroller landscape and settled on AVR, with it's wonderful inexpensive STK-500 programmer/dev board, fully documented specs and protocols, and wide open source support.
Now with CircuitPython, I don't need to deal with programmers, protocols, or any of that. I can program a CircuitPython board with pretty much anything that supports a standard USB serial port.
stuck elbow
the microcontroller landscape has really improved as well
pretty much nobody uses the undocumented and closed-source ones anymore
main meteor
Even TI learned their lesson.
stuck elbow
I'm looking forward to seeing what RISC-V does
main meteor
I'm still cherishing the forlorn hope that there will be open source FPGA support for other than legacy chips.
The Arduino people said they'd democratize FPGA programming, but they're not there yet.
stuck elbow
someone will, eventually
probably by that time everybody switches to proprietary quantum computers, though
it's a red queen race
main meteor
There's a reason I still play with vacuum tubes and relays.
old smelt
@main meteor - I'm interested in a more detailed conversation about open-source FPGA development. Out of scope for this channel, but if you would be open to a 1:1 discussion sometime, I'm all ears.
or we could pick it up in another channel/topic area
main meteor
I'm on travel today so will be in and out
old smelt
Not urgent. Whenever you have some time.
upbeat plover
I've been wanting to learn Verilog, I want a FPGA but the ones Adafruit sell are not quite what I'm looking for. I like the TinyFPGA a lot but I want a little more.
bitter hearth
I'm a full-time Python programmer, and I just started on electronics... I am super duper impressed by CircuitPython's "first five minutes" experience! It's actually way better than CPython's, hahaha 😛 Just plugged in the board, edited the demo file, messed with the code, saw the changes, boom. It's incredible not to have to install any tooling to get to that first step.
orchid basinBOT
prime flower
Thoughts about splitting the helper library list out a bit more into sub-categories? https://circuitpython.readthedocs.io/projects/bundle/en/latest/drivers.html#helper-libraries
such as: Helper Libs -> IoT, APIs (Hue, LIFX, Adafruit IO, Azure IoT), Display (Button, shapes, text, framebuf, image load, slideshow), Testing (board test suite, possibly a unittest down the line), Audio,
raven canopy
@prime flower I personally have no objections. I think it would dovetail with this idea well (for which i've spent a few brain cycles wrt to implementing "easily"): https://github.com/adafruit/Adafruit_CircuitPython_Bundle/issues/140
prime flower
Or just additional organization within that general category similar to how Drivers are organized.
@raven canopy I just clicked 5 helper links at random, and none of them had github topics.
i think we'd need to sweep that list
raven canopy
i'm not sure a single library has one. its not an easy task to update 160+ pages. but...Python is a wonderful too when combined with the GitHub API. 
raven canopy
i only pointed it out since there seems to be some drift in how things are organized/categorized.
prime flower
yup - i agree. if we get this generating automatically, and categorically, we could add a "supported hardware" list to https://circuitpython.org/libraries
which is then search-ible like boards, sort by category
raven canopy
the PyPI ones would be easy, honestly:
keywords = read_setup_py()
github_api_update.topics(keywords)
though, the keywords may not be the best candidates.
prime flower
they're good for specifics but not broad categories, most of the ones i skimmed thru
upbeat plover
@tulip sleet thank you for tip with frozen modules, i used CPX for example and now everything is working as expected
tulip sleet
@upbeat plover Great!
graceful heart
@pastel panther @prime flower Just got home, free to chat whenever
pastel panther
@graceful heart I need to wrap some stuff up and gather some thoughts but I should be good to go at 7
graceful heart
oh ok, cool.
bitter hearth
time.monotonic() is a floating point number of seconds, right? On a trinket m0 express, that's 30 bits? Won't that hit precision barriers within a few months?
ah, I guess people have been thinking about this already, it's mentioned in passing in some github issues
i see, there's monotonic_ns() for some boards
bitter hearth
So if I want to run 10 steps per second on a stepper motor forever, I should use monotonic_ns() and compensate for wrapping around sys.maxsize?
onyx hinge
status: creating audiopwmio in shared-bindings / shared-module and adapting nrf pwm audio to use it
simple pulsar
@bitter hearth In case you've not read it already, time.monotonic() discussed in https://forums.adafruit.com/viewtopic.php?f=60&t=141785
bitter hearth
@simple pulsar thanks, the thread identifies the issue but I don't see it resolved there 🤷
It does point out that I should look at the C source for .monotonic() though...
I only want an accurate dt within a second of the last time I checked, so I think I can just subtract the last common_hal_time_monotonic() from the current one and use that... just have to do it in C instead of python
prime flower
@pastel panther @graceful heart I'm in the vooice chat when you're ready
pastel panther
@prime flower im mid reflow
prime flower
ok!
bitter hearth
Are there CircuitPython C libraries that aren't built as part of the core? I'm clicking around the library bundle and not finding any
guess I should be in the "help-with" channel
onyx hinge
@bitter hearth I don't think you can "dynamically" load a native code module that is not part of the core, like you can import a .pyd on windows python...
bitter hearth
ah, okay, thanks
pastel panther
aahh heck, there goes that
slender iron
@raven canopy what is the right way to run adabot so it adds the environment variables to travis for new repos? python3 -m adabot.circuitpython_libraries -v validate_travis
raven canopy
haha. been too long since i ran it...had to look it up: -t/--token -v "validate_travis"
slender iron
kk
onyx hinge
While I'm looking at audioio/audiocore documentation to appease travis, I noticed a lot of blocks that say Sample must be an `audiocore.WaveFile` or `audiocore.RawSample`.. In general, do these sites also permit a Mixer object?
slender iron
@raven canopy how is that different from ADABOT_GITHUB_ACCESS_TOKEN?
@onyx hinge ya, I think they should
onyx hinge
@slender iron OK, I guess I'll just add audiocore.Mixer to the list everywhere I find those two
oh, hm, can you give a mixer to a second mixer?
raven canopy
@slender iron different? i may have misunderstood your question...
slender iron
@onyx hinge I think so
onyx hinge
It seems like it might be an unusual use, but my expectation from the code is that it would work
slender iron
@raven canopy the flag is confusing me. when true, it prompts for a password
upbeat plover
im having an issue with circuitpython-stage my game doesnt display... I can hear the sounds from the game but all that is displayed is Blinka in top-left of black screen
@stuck elbow is this a know issue or have i done something wrong
slender iron
@upbeat plover what version are you running?
upbeat plover
Adafruit CircuitPython 5.0.0-alpha.0-45-gd99d3bd47-dirty on 2019-07-25; Adafruit Feather M4 Express with samd51j19
slender iron
totally possible I broke it
raven canopy
@slender iron which env var are you wanting adabot to set? -t and validate_travis will set the token for Travis in GitHub for the repo and create the GITHUB_TOKEN on Travis.
slender iron
it -t used for more that setting GITHUB_TOKEN?
raven canopy
nope
onyx hinge
https://circuitpython.readthedocs.io/en/4.x/shared-bindings/audiobusio/I2SOut.html refers to audiobusio.RawSample. Is this wrong?
slender iron
ok, it seems weirdly named to me. I expect something like --prompt-password
@onyx hinge probably
raven canopy
its easily changed... 😄
slender iron
yup, will do
onyx hinge
I wonder if there's a board with I2SOut but not audioio. Now, such a board wouldn't have audiocore.. but it wouldn't have had audioio before, so surely somebody would have noticed it was broken.
raven canopy
oh yeah, now i remember. i used -t/--token because -p/--print exists. otherwise i would've done -p/--prompt.
onyx hinge
but "is" will be "was" soon, nrf will have audiocore + audiopwmio
I think that no, nrf doesn't have audiobusio, but from datasheet it does HAVE both pdm for audio in, and i2s for in AND out
slender iron
@raven canopy kk, will probably just not have a short version
actually leave t as the short
raven canopy
the nrf port on my laptop has audiobusio... PDMIn just doesn't work. 😄
onyx hinge
@raven canopy do you mean you have a work-in-progress ? Or that nrf has it, and I'm overlooking it?
raven canopy
WIP
onyx hinge
OK, I hope that the audiocore stuff doesn't get in your way (on the whole I think it should help)
raven canopy
it shouldn't hurt. and being able to play what my mangled filter is trying to do, instead of copying the REPL output and trying elsewhere will help. hehe
tulip sleet
@slender iron FYI, you know https://github.com/adafruit/circuitpython/pull/2011 has travis failures, right? [moved from wrong channel]
slender iron
nope, will look later
tulip sleet
i assume not urgent due to upcoming weekend
slender iron
makes sense since I changed the constructor C api but not the direct calls to it 😃
nope ssd1322 driver won't work but that's ok
I'm camping all weekend anyway
onyx hinge
'night all. I'll check my pull#2000 for travis errors in the morning, probably I got something wrong while fixing docs or enabling audiocore based on audiobusio too
(I'm just scared of pause/resume and making work to avoid it)
slender iron
night!
onyx hinge
@raven canopy fwiw I ended up dialing back the cleverness of when to enable audiocore, so you'll just want to toss CONFIG_AUDIOCORE = 1 in your .mk after all.
stuck elbow
clever code rarely is
manic glacierBOT
heady raft
anyone around that can help me? I am having problems with getting the bootloader folder to show up on a trinket - i have followed these steps and the usb driver will install and show the usbtiny device but when I double press the reset switch no folder appears on my pc for me to load a uf2 file.. im on windows 7 x64 and also tried a windows 10 pc with the SAME results https://learn.adafruit.com/welcome-to-circuitpython/troubleshooting
manic glacierBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
This was easier than I thought!
lunar crown
I put in a pull request about two weeks ago, fixing some typo class bugs in the Adafruit_CircuitPython_SI5351 driver. It was approved and merged the same day. But, it has never shown up in the Adafruit_CircuitPython_Bundle distribution. Is there something else I need to do to get it into the bundle?
raven canopy
@lunar crown the library needed a new release. i took care of it, and it should be in the bundle by monday at the latest.
meager fog
@heady raft you have a classic tiny85 trinket not a trinket m0, it cannot run circuitpython (its a very old board)
lunar crown
@raven canopy TNX
craggy galleon
Can I just say Great Job who has already ported over the boards to the CircuitPython.org website. I started porting over my board but it is already there 😄
Thank you!
upbeat plover
@gilded cradle I have to round the backlight float or it gets wacky. Without rounding it skips around like it should be 0.96, 0.92, 0.88, ect. but it does stuff like this 0.9645, 0.9189, 0.8548
from minitft_featherwing import wing
import time
# Turn off backlight
wing.backlight = 1.0
time.sleep(0.1)
while True:
for i in range(25): # fade on
wing.backlight = round((wing.backlight - 0.04), 2)
time.sleep(0.05) # slow down signals to seesaw
for i in range(25): # fade off
wing.backlight = round((wing.backlight + 0.04), 2)
time.sleep(0.05) # slow down signals to seesaw
@upbeat plover Is it hurting things that the numbers aren't exact? One alternative would be to write (for the "fade on" case), wing.backlight = i / 24 so that the backlight changes from off (0.00) when i=0, to full on (1.00) when i=24 on the last loop.
(or maybe (i+1) / 25 .. recalling that range() likes zero, so range(25) goes 0, 1, ..., 23, 24, which is 25 numbers, but doesn't actually include 25)
solar whale
@tulip sleet when I update my clone of master I get this error Synchronizing submodule url for 'tools/uf2' Synchronizing submodule url for 'tools/usb_descriptor' error: Server does not allow request for unadvertised object 1dabc2df8ed1acaf0c483b58d757ba72e8b42090 Fetched in submodule path 'lib/tinyusb/hw/mcu/st/stm32lib', but it did not contain 1dabc2df8ed1acaf0c483b58d757ba72e8b42090. Direct fetching of that commit failed. Failed to recurse into submodule path 'lib/tinyusb' do I need to do something different -- I update with this ```git submodule sync
git submodule update --init --recursive
git submodule foreach --recursive 'git fetch --tags'
@solar whale I encountered the same thing recently. tannewt recommended me to "git submodule sync", and I think that fixed it.
solar whale
Thanks -- did not seem to help for me ...
onyx hinge
drat, I was hoping I had the answer
and now that I read what you said, you already had tried that
solar whale
trying a clean clone
tulip sleet
@solar whale I saw similar issues, but a clean clone fixed it
even my super duper:
alias gitsubupdate='git submodule sync --quiet && git submodule update --init --recursive'
did not fix it
solar whale
ok -- happier after clean clone
stuck elbow
obviously we need to add more arguments to that command
maybe a couple of --force and --hard will help?
gilded cradle
@upbeat plover, you could always add a new property or function if that helps. I just didn't want it breaking existing code.
upbeat plover
@marble hornet what was the reason you converted 0-255 to 0.0-1.0? you were first one to do so in https://github.com/hexthat/TFT-FeatherWing-Lib/blob/master/lib/adafruit_mini_tft_featherwing/tft_featherwing.py
It takes longer to execute the code and creates even more math down the line.
@gilded cradle its not even correct the way it is now
wing.backlight = True
turns off the backlight not on
manic glacierBOT
marble hornet
@upbeat plover I used a float because it is closer to a percentage and It felt more intuitive
And how often is brightness set: once or twice?
And while 255 is used for colors that is a paradigm people learn when they start doing display programming but display brightness is a paradigm you many people are familiar with before they start embedded / display work
manic glacierBOT
gilded cradle
@upbeat plover, I wanted it to be between 0-1.0 to be consistent with the neopixel and dotstar brightness values. We could check the parameter type passed in and if it's bool, set the value to 255 (skip the calculation) andFalse to 0. We could potentially take anything over 1 and use it as a raw value and anything between 0 and 1 to be the percentage. Only confusing value would be a value of 1 since it may mean full brightness and may mean nearly dim. That's why I suggested a new property or function.
raven canopy
who wants core docs built with Sphinx 2.1.2 (current)? cause, i have 'em building locally. 😉
sadly, i barely understand how i've done it. 😆
manic glacierBOT
tulip sleet
@raven canopy sounds great! if you want to do a PR that would be fine.
upbeat plover
how do i build older version of CP? im stuck on 5.0.0
stuck elbow
checkout an older tag
raven canopy
@tulip sleet due to changes in an existing PR, its getting tagged on there: https://github.com/adafruit/circuitpython/pull/1985
just battling Travis. the usual "works local, never on CI." game. 😄
also just noticed earlier, Travis finally has Bionic available.
onyx hinge
Because what I always need is another distraction, I was going to look at changing how lis3dh.shake (the method which checks if the device is being shaken) works. Where do I put the file so that it will override the one that is frozen in? (note: can't actually start on this right now, the cpx is at home and I'm still in ohio -- just gathering information)
raven canopy
you can force the import location, iirc
from .lib import lis3dh
can't remember if the . relpath works in cirpy...
onyx hinge
Makes sense, I'll try that first.
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
KeyError: __main__
``` well that doesn't seem happy.Looks like maybe putting it just directly in the top level could do it
raven canopy
try without the .?
onyx hinge
is there a meeting tomorrow? I might be able to at least listen in, if I can find some free wifi at the right hour
raven canopy
should be. i don't remember any cancellation notice.
onyx hinge
I'll try to check in
slender iron
@onyx hinge yup! we're meeting tomorrow
Hi all, we'll have our normal CircuitPython meeting at 11am Pacific / 2pm Eastern tomorrow here on Discord. Everyone is encouraged to join. Here is the meeting notes doc for it: https://docs.google.com/document/d/1Nf7-76XfJN_4uTwfvHyRNtHLUNHfL_-U6BjF6ZzAlGk/edit?usp=sharing <@&356864093652516868>
raven canopy
@slender iron i'm going to merge the "check travis builds new boards" PR, if you don't have anything further? its gonna raise a few conflicts on open PRs, but thats easy to deal with.
slender iron
sounds good. thanks @raven canopy
craggy galleon
@raven canopy how do I go about resolving conflicts about TRAVIS in my pull request no. 2014 ? I would imagine this could be an issue impacting all other boards.
onyx hinge
@craggy galleon According to github help, you can resolve the conflicts using the web interface, which will help by showing you areas that have been changed both by you and other contributors. https://help.github.com/en/articles/resolving-a-merge-conflict-on-github
do you have that button when you look at your pull request?
craggy galleon
@onyx hinge I know where the conflict is. The issue is that every PR for a new board updates the same file (.travis.yml). I am just wondering if there is someone who manages it / resolves them.
Do you have any experience with fixing these types of merge requests?
onyx hinge
@craggy galleon Yes, though I haven't personally done it via the github web interface
do you want to try to walk through the steps together?
craggy galleon
Sure! That would be great!
My concern is that someone else's PR will go before #2014, then the same process would need to be followed.
onyx hinge
I think that your change was just to add a new board or board(s) to those lines, right?
craggy galleon
That's right. 😃
onyx hinge
In that case, the way I would do it personally is to just delete the lines from "<<<" to "===" which represent your changes, and re-add your new board(s) to the TRAVIS_BOARDS lines that remain.
Also delete the ">>>" line
craggy galleon
Yeah sure - that sounds easy enough. Give me a minute and I'll do that now.
Should I double check which boards have changed?
onyx hinge
Well, by starting with the side that was marked "master", you know you are starting with the most up to date list. So I think you can be confident and just add your new board
it is a very simple case, just not quite simple enough for git or github to be able to do it automatically
onyx hinge
OK, let me pull up the pull request again
craggy galleon
Travis CI is just checking I didn't break anything. Thanks Travis! 
onyx hinge
the github page is loading slowly for me, but yeah it looks good
congratulations, you now know how to resolve a merge conflict!
craggy galleon
Thanks @onyx hinge !
onyx hinge
also, thanks for your work on circuitpython. I'm sure somebody will be along soon to (hopefully) approve your contribution!
raven canopy
@onyx hinge @craggy galleon thanks for getting the conflict worked out! it's an unfortunate situation when the PRs get stacked up like that, but its unavoidable. (I was this close to working them myself last night, but I'd rather the PR owners know what's happening so to keep their branches up to date.)
onyx hinge
hmph, trying to connect to the audio meeting from a Panera, but I'm getting "no route". Must not comply with their policy.
onyx hinge
@drowsy geyser can you say something over the voice channel for a minute? I'm trying to figure out whether I can listen this way or whether I need to switch over to tethering
(I won't be talking, but I did put my hug report and progress report in the document)
drowsy geyser
@onyx hinge I can't really - open office space. 😃 I'm ostensibly working.
river quest
newsletter DRAFT for folks who want a sneak peek of what is coming out on tues, and what will be in some of the notes/etc. for the weekly here in a few mins 😃 https://github.com/adafruit/circuitpython-weekly-newsletter/blob/gh-pages/_drafts/2019-07-30-draft.md
onyx hinge
@drowsy geyser I understand how that is
turbid radish
Hi all!
onyx hinge
ah here come some people
stuck elbow
o/
drowsy geyser
Howdy.
stuck elbow
← lurking
old smelt
Lurking.
tidal kiln
lurking
onyx hinge
low power sounds like fun to hack on
drowsy geyser
Lurking today. Hugs, though, to Jerry, LadyAda, and Brentru for analyzing and fixing the "gamepadshift" error on PyPortal with the latest firmware and bundle.
stuck elbow
deep sleep at the least would be super-useful
onyx hinge
years ago, "jeenode" did a blog for years about this sort of thing. he was doing sensors.
hi scott
drowsy geyser
Oh, so I guess I'm text-only instead of lurking. 😃
slender iron
river quest
PyOhio - Kattni's Keynote!
https://twitter.com/EWDurbin/status/1155111965053915137
https://twitter.com/WillingCarol/status/1155165692934574081
Twitter
STM32 support snakes its way to CircuitPython!
https://blog.adafruit.com/2019/07/28/stm32-discovery-f412zg-and-f411re-support-snakes-its-way-to-circuitpython-stmicroelectronics-st_world-circuitpython/
8/8/2019 is CircuitPython day!
https://blog.adafruit.com/2019/07/29/diy-light-painting-wand-using-circuitpython-by-iayanpahwa-circuitpython/
Scott wil be on ASK AN ENGINEER and we have some cool surprises next week!
In the news:
The MagPi Magazine: Control servos with CircuitPython & Raspberry Pi <-- MakerMelissa!
Issue 21 HackSpace magazine: CircuitPython storage
stemma and stemma QT - easy connections!
https://youtu.be/_Cu7UOcGL14
https://learn.adafruit.com/introducing-adafruit-stemma-qt?view=all
tidal kiln
excellent idea!
river quest
Snekboard v0.2 Update: Snek sneks its way to CircuitPython
https://keithp.com/blogs/snekboard-0.2/
onyx hinge
snek is a cool language in its own right
river quest
Made with Mu - A HyperCard inspired GUI framework for beginner developers in Python
https://github.com/ntoll/pypercard
https://twitter.com/ntoll/status/1155567530876055552
Help name our Pygantic PyPortal :)
List of names in the newsletter:
https://github.com/adafruit/circuitpython-weekly-newsletter/blob/gh-pages/_drafts/2019-07-30-draft.md
sterile bronze
just lurking
ionic elk
Whoop I don't know if this is a big deal but we might not want to use Nucleos as the STM32 pics in the news issue! They aren't supported and aren't on the schedule as they lack critical USB components.
We're focusing exclusively on the Discovery line
turbid radish
Speak up Brent
ionic elk
Links if you need em
https://www.st.com/en/evaluation-tools/32f412gdiscovery.html
https://www.st.com/en/evaluation-tools/32f411ediscovery.html
prime flower
@turbid radish Just adjusted my mic - factory is loud 😃
turbid radish
@ionic elk - noted, thank you!!
gilded cradle
I think Kattni is out the next 2 meetings still
onyx hinge
😃
drowsy geyser
@gilded cradle I think so, but I haven't caught up with her after her keynote.
slender iron
@drowsy geyser did you have hug reports to give?
drowsy geyser
@slender iron Yup, hugs to Jerry, LadyAda, and Brentru for fixing issue #46/#47 on PyPortal (the gamepadshift bug).
onyx hinge
what I think I overheard about videos is that everybody organizing pyohio is going to take a deep breath before getting the videos online so be patient
they really knocked it out of the park, they deserve hug reports too
errant grail
"Medicine Man" product claims.
tulip sleet
It's a floor wax and a dessert topping: https://www.nbc.com/saturday-night-live/video/shimmer-floor-wax/n8625 (from 1976)
drowsy geyser
Nothing for me
onyx hinge
@tulip sleet I won't get the paraphrase right, but Greg Svoboda (who gave the other keynote) quipped that when he learned hexadecimal, people told him, "you'll never need to know that". Well, when he learned bluetooth, he needed hexadecimal.
that got shown off at pyohio
hm, the "regular" gcc that is in debian isn't in raspbian? https://packages.debian.org/buster/gcc-arm-none-eabi
drowsy geyser
Thanks, team. Keep up the fantastic work!
errant grail
Thanks everyone!
onyx hinge
so happy to sit in on the meeting today thanks to pyohio and a travel day. I'm going to try to shift my regular schedule to enable me to come more often.
tulip sleet
@onyx hinge arm-none is for bare boards. We'd recommend using the download from the arm site. The ARM toolchain developers seem to no longer be doing any packaging themselves
errant grail
The hardware concept for the Waveform Gen FeatherWing's CPy library/driver
raven canopy
@onyx hinge @tulip sleet I wasn't aware Buster had ver7-q2; it may be an option now. But I doubt they'll do any better at keeping the package rolling with version changes (Stretch is still on ver5). That is why I was following the new "download straight from ARM" approach. They only offer 64bit, or source...
and, I'm still a bit noobish at this level of knowledge
onyx hinge
no, the version in buster won't see updates unless there's an actual security problem discovered. buster-backports might. I understand why the project would standardize on a particular one so that everyone seems the same results as travis, but in practice the debian buster compiler has been working for my nrf stuff
and then the part where I'm noobish is how / when things get from debian into raspbian, sooo
raven canopy
Looking at that package, they have an armh version, so it should get rolling released on the RPi.
I'm still not sure that arm64 ( Debian renamed from aarch64) will run it. I'll most likely be learning/fighting with dpkg. 😁
errant grail
@slender iron Looks like @simple pulsar is already aware of the MIDI running status issue. https://github.com/adafruit/Adafruit_CircuitPython_MIDI/issues/8
upbeat plover
something changed in audioio that broke stage
File "stage.py", line 166, in play AttributeError: 'module' object has no attribute 'WaveFile'
stuck elbow
@upbeat plover pull request please? :)
upbeat plover
@stuck elbow PR done almost did typo and do "audioiocore" but i caught it and did "audiocore"
solar whale
should the audioio/core changes be made so thay are "backward compatible" try/except?
upbeat plover
im still having issue where ugame stuff doesnt show, just blinka top-left and black-screen, everything works fine 4.1.0-rc.1
stuck elbow
@upbeat plover is that an nrf?
upbeat plover
feather_m4_express with minitft_featherwing
@solar whale why do i need both audioio and audiocore? AudioOut is only in audioio
@stuck elbow the PR has the sound working now, but now both audioio and audiocore are imported
slender iron
audiocore contains the source of sounds now (aka RawSample, WaveFile and Mixer)
upbeat plover
"When nrf pwm audio is introduced, it will be called audiopwmio. To enable code sharing with the existing (dac-based) audioio, factor the sample and mixer types to audiocore." Is this so if you dont need everything you can just import AudioOut if thats all you need?
slender iron
it's so that ports can support audiocore without audioio. (the nrf doesn't have a dac)
tawny creek
random Q, would the adafruit_is31fl3731 benefit from being part of displayio?
slender iron
it'd certainly be interesting!
tawny creek
lotsa similarities to the oled stuff
slender iron
16x9 isn't a lot of pixels
tawny creek
not at all, would the fonts have to be redrawn?
slender iron
the builtin is smaller than that but it wouldn't fit much text
I kinda assume we'll want some sort of scroll for displays that small including the charlcds
upbeat plover
i did PR for CPX so it works with 5.0.0 how do you sperate stuff for like a 5.x bundle?
slender iron
@tulip sleet https://github.com/adafruit/circuitpython/pull/2011 is ready for review when you have a chance
tulip sleet
@slender iron Great! going out soon but will review later tonight or tomorrow morning
slender iron
k, np
onyx hinge
@slender iron it's a perfect time to nitpick me, I'm at the airport and it's an alternative to looking at flight delays
🛫
manic glacierBOT
slender iron
👍
tawny creek
or a very low res game
onyx hinge
oh travis you're so very particular
@slender iron this documentation building problem was caused by the rename; is "one more rename" (so the end user will refer to audiopwmio.PWMAudioOut) necessary, or what do you see as the resolution? /home/travis/build/jepler/circuitpython/shared-bindings/audiopwmio/__init__.c:12:toctree contains reference to nonexisting document u'shared-bindings/audiopwmio/AudioOut'
it looks like an assumption about class names vs C source filenames to me
raven canopy
that's going to get tricky even if it builds, since shared-bindings/audioio/AudioOut.c also uses the class name AudioOut. RTD doesn't have "branching" for the ports...
wait. nevermind. thought just completed. they'll live under different modules...
needs to make dinner so brain functions properly
slender iron
@onyx hinge make sure to pull changes. I tried to fix the last bits
raven canopy
egads... that table has no wrapping. 😬
found fix #1! hehe
but, it all still works as intended, so i'll take that. 🎉
onyx hinge
@slender iron thanks, I'll try to resort the urge to force push the branch
solar whale
@slender iron is there an update to the ssd1306 lib for displayio? Or did I misunderstand? I don’t see a recent update.
onyx hinge
I wonder if there's a way to generate the board list within a a Travis job instead of having to manually maintain it
Even if the balance was not quite ideal
raven canopy
i had that too. even thought about having a log file in S3 to handle load balancing.
onyx hinge
Ooh that would be way beyond anything I had thought of
raven canopy
i mean, at 700+ assets to build, its not a small task we're asking of Travis. its going to take time.
i plan on deep diving on why the Polish builds take like 8000% more time to build. i'm thinking its more to do with its completeness as a translation, but it'd be interesting to try and trim the PL build times. (that % number is wholly hyperbolic...)
onyx hinge
Yeah in that light 50 minutes is pretty fast
Were you involved in making the translations work?
raven canopy
i was not. i don't even understand it fully. 😄
onyx hinge
I haven't looked but want to someday to understand it
Our else join you in perplexity
raven canopy
i think its just a matter of inline-ing from the .po reference files. strong think though.
onyx hinge
As in it literally makes new source files?
river quest
Sign up for the next newsletter, shipping in less than 24 hours - Celebrating 2 years, and 6,000 subscribers! Sign up now "Python for microcontrollers" https://www.adafruitdaily.com
raven canopy
@onyx hinge that's the point where i lose understanding. hehe
slender iron
@solar whale it's a separate library so rpi users aren't stranded
onyx hinge
it looks like the translation procedure produces a very big function which linearly searches through the list of all original strings and returns the related translation. It must be depending on the optimizer to find at compile time which one of the strings is actually the result. However, it seems like the size of the translation is always the same no matter the target language
onyx hinge
.. in translate.c, inline __attribute__((always_inline)) const compressed_string_t* translate(const char* original)
@pulsar ferry were you serious about "pt" translation taking longer to build? I didn't duplicate that finding here, not even 10%
though I'm not using the sanctioned gcc version, I'm using whatever's in debian buster
raven canopy
@onyx hinge not pt. pl. its always around a minute. e.g.: pca10056 takes <3 seconds for every language, and pl takes 52 seconds.
though, i just realized that its the first language to get built for each board. i vaguely remember a "don't rewrite po files" speedup recently.
hehe
discord was autocorrecting me. hilarious.
manic glacierBOT
could be related, but i was seeing 0x3 in chstatus which would be BUSY + PEND
raven canopy
only when reading other people's autocorrect misadventures.
onyx hinge
.. and now bad airport wifi is about to cut me off unless I pay
so bad it made me say a cuss, sorry 😦 😦
goodnight @raven canopy thanks for encouraging me to understand how CP does translations a little better!
raven canopy
@onyx hinge night! enjoy the air travels!
raven canopy
that's the best part. only good part? something like that. 😄
onyx hinge
it is amazing that when it works you can go all the way across this huge country in just a fraction of a day
manic glacierBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
prime flower
Is binascii not in within CircuitPython 4.x builds?
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
ImportError: no module named 'binascii'```same with hashlib too even though it's listed on latest? https://circuitpython.readthedocs.io/en/latest/docs/library/hashlib.html
prime flower
oh, hrm
Some libraries are only enabled only WiFi-capable ports (ESP8266, nRF) because they are typically used for network software: binascii, hashlib, uheapq, uselect, ussl.
prime flower
Are there plans to add these to libraries to 5.x since WiFi /networking is utilized by some boards (AirLift, PyPortal)?
manic glacierBOT
orchid basinBOT
onyx hinge
I'm getting an unusual build error. ```Warning, treated as error:
failed to reach any of the inventories with the following issues:
intersphinx inventory 'https://circuitpython.readthedocs.io/projects/busdevice/en/latest/objects.inv' not fetchable due to <class 'requests.exceptions.SSLError'>: HTTPSConnectionPool(host='circuitpython.readthedocs.io', port=443): Max retries exceeded with url: /projects/busdevice/en/latest/objects.inv (Caused by SSLError(SSLError(1, '[SSL: UNSUPPORTED_PROTOCOL] unsupported protocol (_ssl.c:1056)')))
I'll try again in a bit in case it's a problem on their end, but circuitpython.readthedocs.io loads in my browser
sphinx-build 1.8.5 running with python3 (everything from pip), and I swear it worked last week
raven canopy
@onyx hinge with my PR getting merged yesterday, we're now running on Sphinx 2.1.2. but that looks...Travis-side.
onyx hinge
@raven canopy huh so despite trying to use pip(3) to get it, is my sphinx seriously out of date?
smh
raven canopy
shouldn't be too out of date. if you do update to sphinx 2.x (py3 only), you'll need to pull from master though, as there is an update to c2rst that allows it to work with 2.x.
brazen bluff
Hello I'm brand new to the circuitpython community. I had the distinct pleasure of Meeting @idle owl at PyOhio and she encouraged me to begin my journey by joining this community. I appreciate all her words of encouragement and I'm excited to be here and start on my GBPhoneHome Project (name credit goes to Dan Lindeman)
raven canopy
👋 welcome @brazen bluff!
onyx hinge
@raven canopy it's working now, so yay?
onyx hinge
hmmm looking at the support matrix, which is my goal, I see that there are some boards that have "audiocore" but not "audioio", like Adafruit Feather M0 Adalogger. This may be a mistake due to my PR that added audiocore.
@brazen bluff welcome! I was at pyohio but I didn't meet anybody who looks like your avatar there. Don't miss the #help-with-circuitpython channel too
brazen bluff
@onyx hinge now that you mention it I didn't see any owls ? 😃
onyx hinge
🦉
@raven canopy yeah I did something "smart" for audiocore that confuses the autogenerated support matrix. I think it's not actually the case that "audiocore" is in the Adalogger, just that the script doesn't understand what I put in the .mk file
raven canopy
the support matrix is fresh...and rigid. i wouldn't make decisions based on its result. its easily changed... 😄
onyx hinge
here's what I did, which was intended to make it so that boards with audioio automatically got audiocore. It was lazy.
raven canopy
ahh. there are a few others like that already. wiznet, terminalio, etc.
i would've brought it up #in-the-weeds, but i couldn't attend the meeting. just need to have a larger conversation about it. one option is to just write the special cases into support matrix. core code shouldn't exactly bend to the docs.. 😄
onyx hinge
Would you like me to file an issue, or do you think you'll remember this?
raven canopy
i'll remember. its got a wrapping bug. but, the initial issue is still open, so i'll add it there too. (i noted it in the PR, as well)
onyx hinge
OK, thank you for working on this!
raven canopy
doh! i missed a tannewt reply to the original issue recommending closure. new issue it is. hehe
onyx hinge
make -n -p BOARD=... | grep ^CIRCUITPY_ will get you the content of relevant makefile variables after all the parsing is done, but sometimes variables still refer to other variables in this listing.
near wedge
Anyone know if you can edit code on a CircuitPython board using an iPad and USB connection kit?
raven canopy
@onyx hinge yeah, that's the issue at hand. the script reads cirpy_mpconfig.mk to establish the "base", then reads mpconfigport.mk and mpconfigboard.mk to determine inclusion/exclusion. but for those few that don't have a direct on/off correlation, they aren't captured. well, except for FULL or SMALL build.
river quest
raven canopy
was that just really fortuitous timing? lol
raven canopy
@near wedge see blog post ^^^ 😄
river quest
that is kinda cool
raven canopy
indeed!
near wedge
Yay! This will come in handy for Defcon! 😄
pastel panther
@graceful heart I just had a chance to finally try out the color picker server demo. It works great! (once I fixed a bug that had nothing to do with your code) 😃
manic glacierBOT
raven canopy
rolling through the labels for that issue, i realized Hacktober is like 2 months away. 🎃
lone sandalBOT
graceful heart
@pastel panther :)
I think we might want to re write the demo app to not use 3rd party scripts hosted on a cdn. Works great when your client is connected to internet, but I tried using in conjunction with AP and obviously since my device is connected to the esp AP, it doesn't have internet to download the scripts from a cdn...
So we can either include the downloaded assets in the static folder, or just re write the JS app to not need them. Wouldn't be as cool if you had to type in your rgb values though....
manic glacierBOT
slender iron
@umbral dagger ask here please
umbral dagger
@slender iron I'm using fill_area to grab pixels from the display. It works as expected on a PyPortal, but on a PyGamer the result is rotated 90 degrees counter clockwise.
slender iron
the displays are oriented differently
in the init code
pyportal is the weird one
the intention is for the buffer to align with the real orientation of the display
manic glacierBOT
umbral dagger
@slender iron Is there a way to determine the rotation?
umbral dagger
I'm happy with saving a screenshot rotated, but the width and height don't reflex the rotation, so they'll need to be swapped. Knowing what the hardware is could be used to flag special cases... but that's cumbersome as new boards are released.
slender iron
ya, width and height are weird like that
feel free to add a rotation property
could even add the ability to set it
manic glacierBOT
umbral dagger
@slender iron OK... looking into that where would you suggest it? board? display?
I guess Display
slender iron
@umbral dagger yup, display
@tulip sleet @raven canopy is the latest mixer code here: https://github.com/adafruit/circuitpython/pull/1365? (@onyx hinge for context) I restored the branch because it had been deleted
raven canopy
@slender iron I believe that is where I left it. I still have my local branch. I can get it back to current an submit a PR.
prime flower
@slender iron Would it be possible for binascii to be built with 4.x? I tried adding it to mpboardconfig.h and got 3 of errors relating to mbedtls/ . The pure-python implementation I have works but it's pretty slow
slender iron
@prime flower what do you mean by possible?
@tidal kiln any idea why usb would fail on one pybadge but not another? https://forums.adafruit.com/viewtopic.php?f=60&t=154635
prime flower
@slender iron I wanted to import binascii (https://circuitpython.readthedocs.io/en/latest/docs/library/binascii.html) so I can use it with a RSA lib I'm working on. But I found out that it's only enabled on wifi-capable ports (ESP, which doesnt exist in 4.x, and NRF).
Are there plans to re-enable it on samd ports in the future, or would I be better off going with a python implementation?
slender iron
I don't have plans to add it but that doesn't mean it couldn't be added. What APIs of binascii do you need?
tidal kiln
@slender iron weird one. not sure. want me to iterate a bit to offload you?
prime flower
slender iron
@tidal kiln sure! I'm not sure what the next steps are with them.
prime flower
sha256 would also be useful from hashlib, as well. (https://circuitpython.readthedocs.io/en/latest/docs/library/hashlib.html#hashlib.hashlib.sha256)
slender iron
I'd suggest filing issues for binascii and hashlib.
prime flower
Sure!
slender iron
binascii should be moved to the shared-bindings/shared-module split
solar whale
@graceful heart I finally got around to trying the wsgiserverdemo. It worked great -- so much to learn... Thank you!
graceful heart
@solar whale 😃 
umbral dagger
@slender iron PR soon
upbeat plover
@slender iron i did
try: import audiocore except ImportError: import audioio as audiocore
but it still is failing travis... I tested code on my CPX and my test melody.py is playing on stable 4.0.2 and 5.0.0
manic glacierBOT
raven canopy
@upbeat plover it looks like pylint is complaining about audioio possibly being imported twice:
import audioio
try:
import audiocore
except ImportError:
import audioio as audiocore
This is a tricky one... 🤔
stuck elbow
you can replace import audioio as audiocore with just audiocore = audioio
raven canopy
That seems like a useful trick here. I do wonder how mangled the global is after that ImportError. It may work, but imagine it's an interesting web...
stuck elbow
not sure what you mean by mangled?
ImportError happens before any code is executed from the module being imported
raven canopy
Right. But audioio is already imported...
stuck elbow
looks like new displayio broke stage :(
upbeat plover
yeah i havent been able to get ugame to display in 5.0.0
stuck elbow
@raven canopy imports are cached, you can re-import a module as many times as you want, and nothing will happen
I wish I weren't so sick, I could look into it
@slender iron any idea what has changed? I basically only call displayio_display_set_region_to_update and display->send
stuck elbow
I can't see anything suspicious in those functions
manic glacierBOT
onyx hinge
@upbeat plover @raven canopy I sure didn't try that compatibility block under pylint. It sounds like the assignment version mentioned by @stuck elbow might be a workaround
onyx hinge
If you came to a conclusion about what did work, I'll get it added in a note on that PR
(it's confusing, at $DAY_JOB, PR is a problem report, so sometimes I say PR when I mean issue..)
raven canopy
the other option is to just disable the pylint flags.. 😉
onyx hinge
@raven canopy with the support matrix, was it "just coincidence" that it got transposed (so that boards run down the left and modules go across), or was it a deliberate choice?
raven canopy
this is going to be an interesting merge conflict unravelling on my Mixer branch. good thing compilers are smarter than i am, and can point me in the right direction. 😄
stuck elbow
fixed it
raven canopy
@onyx hinge deliberate choice.
onyx hinge
I still am thrilled every time I use a debugger on a microcontroller. it's like the future
OK, time to figure out how to pause/resume audio! or at least spend a solid 30 - 45 minutes working towards taht
raven canopy
@onyx hinge it does look much better when the text wraps properly...
onyx hinge
oh yeah you are on the right track
and it will be so much better for humans that they can look up the board name, not have to know if it's a samd
stuck elbow
by the way, I see there is meowbit in the adafruit shop, any chances for having circuitpython running on it? it's stm32f4 wth an extra flash spi flash
upbeat plover
i was just looking at that, cool new board
stuck elbow
it's been around for a while
their support sucks hard, though
they have a license-breaking closed-source fork of the mu editor executable, only for windows, for example
(mu is licensed GPL, they don't provide source code)
upbeat plover
when i was first looking at a dev board i was gana get a micro:bit, but i heard they were UK made and i seen CPX was USA... now im here
onyx hinge
alright! pausing and resuming a RawSample now works on nrf
and a wave file
there was just a bug in my wave file test program 🤕
raven canopy
just a normal day in software land. build 3 just failed on bringing MixerVoice up-to-date.. 😄
onyx hinge
Are you getting further each time though?
I'm going to go do a happy dance for a little while and then re-test everything and find out what I broke while adding pause...
discovered and fixed one bug in here, though: a RawSample would only loop, I think 131070 times...
raven canopy
haha. clearly not enough. "Loop Uptime: 4 years, 16 days, 13 hours, 6 minutes, 28 seconds"... 😄
onyx hinge
I'm using a sample program that makes one repetition of a "440 Hz" sine wave, so that's right around 5 minutes
actually testable
onyx hinge
travis is having trouble counting how long things take today
was it really fast? really slow? who can say!
@tulip sleet fwiw I'm looking at the new gcc snapshot, at the point of seeing if the travis build will finish. before_script.19 388 0.01s$ (! var_search "${TRAVIS_SDK-}" arm || arm-none-eabi-gcc --version) 389 arm-none-eabi-gcc (GNU Tools for Arm Embedded Processors 8-2019-q3-update) 8.3.1 20190703 (release) [gcc-8-branch revision 273027]
on the off chance it succeeds I'll put up a PR; if it fails, I'll add notes to https://github.com/adafruit/circuitpython/issues/1887
onyx hinge
blargh ```Old compiler (7-2018-q3-update)
Build trinket_m0_haxpress for pl took 72.58s and succeeded
Build trinket_m0_haxpress for zh_Latn_pinyin took 18.23s and succeeded
https://travis-ci.com/adafruit/circuitpython/jobs/221025799
New compiler (8-2019-q3-update)
Build trinket_m0_haxpress for pl took 199.69s and succeeded
Build trinket_m0_haxpress for zh_Latn_pinyin took 61.75s and succeeded
https://travis-ci.org/jepler/circuitpython/jobs/566215443
that's not gonna fly
tulip sleet
@onyx hinge - yeah we did this experiment a few weeks ago, and we were really disappointed
maybe due to -flto
onyx hinge
@tulip sleet I don't think things will work without lto, will they?
tulip sleet
they'll be too big. i need to do some websearching to see whether there's a fix for this (like turning off some expensive optimizations or something)
onyx hinge
I'll document my findings so that a third person doesn't come along and try the same thing
What a disappointment what such a huge performance regression could go out. I wonder what the story is.
It maybe they just built the compiler with optimization off
tulip sleet
try an m4 build without -flto and see if gcc8 is closer to gcc7 there.
onyx hinge
I can do that while I'm documenting timings, sure
Supposedly clang can build for arm. Has anyone tried that, that you know of?
Hmm the build failed anyway. I didn't drill down yet to see why
Oh, time limits
😭
slender iron
@onyx hinge I tried clang a while back and the code size was 10s of kilobytes larger
onyx hinge
@slender iron 😦 okay
slender iron
ya, I'd love if it worked
onyx hinge
hm, any timings I generate locally are suspect since my baseline is debian stretch's gcc, not the one y'all use
(same for sizes)
stuck elbow
one of these days someone will rewrite circuitpython in rust
slender iron
I'd love it if someone figured out how to interop with rust
onyx hinge
yeah I heard that a wealthy recluse with a lair on a volcanic island was going to provide the funds
slender iron
would love to use it for new modules
stuck elbow
I think it generates normal object files?
slender iron
¯_(ツ)_/¯
stuck elbow
@onyx hinge in my experience funds don't help much, a bored programmer helps much better
upbeat plover
@stuck elbow your stage audiocore changes havent made it into current CP
with fixes, everything working for feather_m4_minitft_featherwing the ugame
exotic pumice
rust has good ffi
probably not impossible
do you need c-interop or python interop or both?
@slender iron
and c-interop is even easier
graceful heart
hi all! Question for everyone: What do you think is the best way to author a project that has dependencies on Circuit Python libraries (like from the bundle, for example) ? Does a requirements.txt make sense? I don't think a pip install would download the deps in an easy to get to location though?
Is the best option just to write a well documented Readme that points to the latest release of each library the project has a dependency on?
exotic pumice
this looks good too https://crates.io/crates/pyo3
raven canopy
@graceful heart yeah, at this time, pointing to the latest release for either the bundle or the library itself is best. for non-Github-oriented folks, this is probably the best link: https://circuitpython.org/libraries
takes a little guesswork out.
minor plume
@graceful heart I was working a little bit on a python script that would look at a yaml config file and copy listed circuit python libraries from the bundle onto a device. I stopped working on it for a bit because I realize I was making it over complicated, but I may finish it at some point just in case people want to use it.
https://github.com/tammymakesthings/cpydeploy has my code as I left it when I realized I was making it more complex than it needed to be.
raven canopy
hehe. i have a similar one; auto forking and local git updating for every library repo. GitHub API is a wonderous thing. 😄
minor plume
I need to keep reminding myself of that principle of “do the simplest thing that could work, and add complexity when you need it and not before” 😂
Premature optimization is a real problem sometimes.
raven canopy
but that's just...no fun. haha. i like cpydeploy; looks good!
minor plume
Thanks! I need to get back to it and at least clean it up. I also need/want to start contributing to circuit python, and finish a couple of projects I have on my bench. However, I am moving from CA to AZ in a month, and I suspect none of those things is going to happen until after I move.
raven canopy
yeah, moving definitely stalls other stuff.
minor plume
Yup. I am moving into two new roles at work, presenting a panel at my company‘s customer conference in two weeks, and then moving in early September. A lot going on right now.
I said to a coworker the other day that I’m actually kind of grateful that I have way too much to do to be able to spend time freaking out about how much I have to do. 😂
raven canopy
haha! been there. congrats and good luck on the new roles (if they're desired...)!
minor plume
Thanks! They are, and they potentially pave the way to some very interesting opportunities a year or so from now.
raven canopy
its "break time" for me. i over-pruned this merge conflict, and my brain is telling me "nope. not doing that right now". 👋
minor plume
Plus they give me an opportunity to really make a difference for my company and for our customers. (I am a recovering DBA/Dev, and a manager at a software company.)
raven canopy
recovering DBA/Dev
🤣
minor plume
Your brain sounds like it’s making a wise suggestion.
I was a decent DBA, and I was a decent Dev before that, but I’m discovering that I really have a knack for leadership. The team I’m going to be leaving to take my new role is in a healthier place than it has ever been at any time in the company’s history, according to my boss.
raven canopy
"proof is in the pudding". leadership isn't easy. and to paraphrase something i picked up somewhere (and use a lot): "management is not leadership. anyone can manage. leading takes passion."
minor plume
Yup. And a fine balance between confidence and humility, I think.
raven canopy
definitely. i've enjoyed most of my leadership roles. some...are just no-win situations. 😄 ok..now i'm off.
minor plume
Yeah.
manic glacierBOT
onyx hinge
Can someone point me at the docs for translators? Or, really, my question is: are translators intended to edit the .po files by hand, or is there some suggested technology to help them out? Or, really, my statement is, I'm excited about https://weblate.org/en/ and I wonder whether it should be investigated for circuitpython
tulip sleet
@onyx hinge only https://github.com/adafruit/circuitpython/issues/1098 There are GUI tools that help manage the editing of .po files, but I haven't tried them
it looks interesting, the free tier for libre projects has some limitations but I don't understand what they are
manic glacierBOT
upbeat plover
anyone know of any other libs that need audiocore fix? I just did CPX with @stuck elbow fix.
stuck elbow
@upbeat plover maybe the rtttl?
upbeat plover
did PR for rtttl
upbeat plover
@stuck elbow not sure why but CP is using "circuipython-stage" 6d1ae72916 and not master
oh i see why in CP it says its using "circuitpython-stage @ 6d1ae72916"
should it be @ master?
raven canopy
Submodules are tagged to commits. On a local, you can point the submodule to master. But the moment you push it, it will reflect the current commit.
upbeat plover
@raven canopy wouldnt that be a good thing? error will be caught faster?
stuck elbow
@upbeat plover no, it uses an explicitly released and tagged version
this way we can keep doing development on the master and not break CP
ruby atlas
@idle owl Welcome back, and awesome keynote at PyOhio!
idle owl
@ruby atlas Thank you!
meager fog
@tidal kiln p0ng
swift arrow
anyone using Visual Studio code to program their CP stuff?
I seem to remember a blog post or something that there was an extension for VSC that enabled the output to be seen.. ie print statements.. but I can't find the extension nor the post.. like the serial console in Mu
raven canopy
@swift arrow this one perhaps? https://blog.adafruit.com/2019/01/02/using-visual-studio-code-to-program-circuitpython-with-an-adafruit-neotrellis-m4-adafruit-neotrellis-circuitpython/
swift arrow
no that is the one that keeps popping up in my searches.. I think it was like an adafruit blog or something
lol which that is also a blog.. but I'm after the one that talks about the name of the experimental serial console like the one in Mu
I wonder if I can't find it because "Mr Lady Ada" talked about it on a show..
argh.. I swear it exists somewhere
raven canopy
The one from Scott Hanselman shows the serial console being used in one of the screenshots. 🤷
swift arrow
doh... nope
tidal kiln
@meager fog p1ng
ruby atlas
@idle owl do you have a dotstar strip someplace I can test the _pixelbuf refactor on? I think I thoroughly broke it in my latest branch.
meager fog
@tidal kiln hihi for the seesaw firmware
its almost certainly that bug
so question - if we dont use the callback to query the # of pixels, it still works
idle owl
@ruby atlas Yep! On it. 😃
tidal kiln
not sure. i think the number of pixels is needed. it's parameter to the call that actually does the writing.
me->m_pixelDataSize here:
https://github.com/adafruit/seesaw/blob/master/source/Neopixel.cpp#L219
ruby atlas
@idle owl Thanks! That strip will do fine. Now to wire it up to the Metro M4 so I can also connect the Segger.
meager fog
@tidal kiln ahh i mean, is it that we only store 255 for the length of the buffer, max, or that we can only address
its def a bug, i think it was a miscommunication, we should support 255+ 'pixels' but got turned into '255 bytes'
the samd09 doesnt have a ton of memory but should be able to do maybe 100 pixels
the samd21 has 'infinite' ram 😃 can do 1000
raven canopy
@ruby atlas I usually use the little grabbers on exposed jumper. can be a pain keeping them in place though..
tidal kiln
the actual buffer size looks like it is set differently for different targets
it's just that variable for the length
ruby atlas
@raven canopy yeah it's a real pain. I need to solder things up more permanently at some point. I've been using alligators.
meager fog
@tidal kiln ok how about we return 16 bit no matter what, then?
tidal kiln
yep. that's the fix.
meager fog
ok please patch it and try it on raspi/feather if you can
im a lil swamped today with a script writeup
but i can also test
tidal kiln
basically just change to uint16_t here:
https://github.com/adafruit/seesaw/blob/master/include/event.h#L479