and i think we even have older 2.x style % formatting examples floating around
#circuitpython-dev
tidal kiln
idle owl
Yeah that is true, and those need to go.
But that's another story.
I'll put Dylan on it.
Thank you for catching this.
For future reference, my suggestion would be to merge a failing PR, and notify me about the issue so we can make a decision without desyncing library config files.
You should have the ability to merge a failing PR.
tidal kiln
yep. i considered that.
blissful pollen
I noticed that I could merge the PR even with the failure but wasn't sure if that was a good idea or not.
idle owl
I noticed that I could merge the PR even with the failure but wasn't sure if tha...
In 99 percent of cases, it is a terrible idea.
In this particular case, where something remote has changed and it's nothing to do with us or your PR, I'm ok with merging and then dealing with the problem. But this is a rare case.
tidal kiln
@blissful pollen yah. maybe you could do that with the PR you're dealing with? otherwise you're just repeating my no-no. ๐
idle owl
@blissful pollen @tidal kiln Repeat the no-no this time. Because we're about to do it globally.
And Adabot knows how to skip things.
blissful pollen
The author did change it already - just haven't merged it yet they have to (last I looked still fix the examples) and the change to this PR has to sync with the equivalent busdevice core change
tidal kiln
okei doke. thanks. @idle owl for future ref - OK to ping you for any CI thing? or is it just a subset?
idle owl
okei doke. thanks. <@!330227457296957440> for future ref - OK to ping you for an...
I'm great to start with, and I usually know who to direct to if it's not for me.
So ping me.
tidal kiln
excellent. thanks!
idle owl
Thank you for catching it!
idle owl
Just mistyped "origin" as "origina" and now I want an Orangina.
mental nexus
@onyx hinge looked through ParallelBus code and looks good to me with what I can see by a quick look at the code. With no hardware handy right now, I did not test. Since I didnโt do testing, I was hesitant to โapproveโ but let me know if you want me to.
onyx hinge
<@!426200547024961539> looked through ParallelBus code and looks good to me with...
@blissful pollen do you think you'd have a chance to do any testing?
@mental nexus appreciatae you taking the time!
blissful pollen
<@!228322543507865601> do you think you'd have a chance to do any testing?
I have not, I may have missed the PR going through which one is it?
onyx hinge
https://github.com/adafruit/circuitpython/pull/5378#pullrequestreview-762377697 I just requested your review
blissful pollen
okay thanks i'll try to take a look at it later today or tomorrow
idle owl
Any CircuitPython Librarians up for approving a few PRs? I have four outstanding ones I'd like to get merged. I can link if anyone is up for it.
Thank you @tulip sleet!
rugged blaze
It would be useful to have a few sentences, and/or a link to more details for each bullet point on CircuitPython release announcements.
I do not have any idea what some of these features are/do (some of the feature names use jargon, and I don't frequent this Discord).
I'd like to have a brief explanation right there in the announcement text that would inform me as to whether I should investigate a new or updated feature further.
idle owl
It would be useful to have a few sentences, and/or a link to more details for ea...
I think the best we can do is consider adding a link to the relevant PR. The releases and their associated notes already take a significant amount of time and effort (multiple hours). Perhaps the PR and its discussion would at least be somewhat helpful to you. @tulip sleet
rugged blaze
Yep, I get that it's a lot of work, and I'm grateful. Of course it's not just me that probably would benefit from this suggestion, but rather your entire target audience.
idle owl
The PR usually has some kind of details. Since... we need to know what we're merging, etc.
We'll consider what we can do. Thank you for the suggestion!
idle owl
@lone axle Thank you so much for the git help yesterday. It worked perfectly. Though I realised as I was typing this, I did not do the other search I needed to do. So I only half completed the job. Oops, I guess it's another opportunity to get the commands into my head.
I searched for wheel(), I did not search for instances of _pixelbuf imports. Fingers crossed I didn't add any, but I bet I did.
lone axle
Hooray, you're welcome. I still don't have it fully commited to memory yet. Pycharm saves terminal histories seperately and I know I've run that in the bundle repo before so I go there and up arrow through my history to find it ๐
idle owl
Yeah, I use iTerm and have some bash magic going on where I can type the beginning of something and up arrow and it goes through all the commands that start with what I typed.
SO convenient.
I'd be lost without it.
Or hitting a lot of up arrows to go through everything. ๐
I didn't know that about PyCharm though.
lone axle
as a heads up the ; true part of that command may not be necessary with all commands. It's needed because git grep returns an error code if it doesn't find anythign which prevents the foreach loop from continuing. So commands that don't behave that way shouldn't need it. Though I don't think there would be any harm in having it even when not strictly neccessary.
ember iris
history | grep 'git' is a good "what was that command again?" trick too
idle owl
Oh good to know.
Thanks.
Man, populating the bundle takes ages.
I delete it every time so I'm not having it all sitting there.
lone axle
`history | grep 'git'` is a good "what was that command again?" trick too
nice! thanks for the tip. That seems to show history from all terminal sessions as well which is cool if I'm in one that has it's own history.
idle owl
Oof, yep, I did miss some.
Luckily not any of the same libs I already did. That would have been embarrassing. ๐
ember iris
nice! thanks for the tip. That seems to show history from all terminal sessions ...
I think* only of past closed sessions unless you've got a 'share the history' setting enabled. So if you've got two terminals, type a command in one, I don't think it show's up with history | grep "keyword". But I don't know how the history thing works sooo milage may vary?
lone axle
I'm not sure if there is some setting for it. I've never changed anything that I can recall but the terminals inside of PyCharm all remember things that I've run inside of each project and terminal instance specifically and don't "up arrow" through commands that were run elsewhere.
ah, running history | grep whatever inside those does seem limited to only stuff that I've run inside of it as well. I must have run that foreach inside of a "normal" terminal at some point because I found it when I first ran that after you mentioned it.
ember iris
Ok, that makes a lot of sense. I was really confused at how much shared history there was, but I don't run terminals in IDEs so it's a major grey area to me
lone axle
Thank goodness for working primarily on a laptop. 
just flickered but came back super fast and battery kept everything in place.
ember iris
oof that's fortunate that the power is back and the battery kept things for disappearing. Time for a round of "Save save save" sprints
idle owl
It's been raining here for 48 hours straight. Had power flickers last night.
Wind died down today which was likely the cause.
blissful pollen
I need to get some trees down a couple years ago after an early ice storm my lights kept dimming and I could see branches banging into my power lines
idle owl
Yeah we have a giant tree in the back yard we keep an eye on. We wouldn't take the whole thing down if it came to it, unless it died. In which case it's a hazard to the house as well as the power lines.
Turns out you have to push if you want your changes to show up on GitHub ๐คฆ๐ปโโ๏ธ
blissful pollen
i have a couple trees that died as result of some sort of insects I think. Some tree service is going to make a nice amount
idle owl
Oi, that's true.
Hey folks, I have two PRs that could use a review. https://github.com/adafruit/Adafruit_CircuitPython_NeoKey/pull/3 and https://github.com/adafruit/Adafruit_CircuitPython_LED_Animation/pull/82
Thank you!
blissful pollen
Oops I didn't notice it but the checks failed on the LED animation
idle owl
Ah have some things to fix on that one.
Yeah I just saw that.
No worries.
Sphinx biting me.
blissful pollen
My git knowledge is lacking for PR #5383 it now has to be merged after microdevs change for the ESP IDF key. Does anything have to be done to do that? I tried to re-run the workflow and it still failed
tulip sleet
My git knowledge is lacking for PR #5383 it now has to be merged after microdevs...
you have to push a nothing commit to get it to redo the pr against the tip
the PR is set to the commit that existed when you did the PR, until you do another commit on the PR
blissful pollen
Okay thanks I'll let the author know
tulip sleet
Okay thanks I'll let the author know
I can merge it anyway with the failures
I thought it needed re-testing, but it's fine except for the bad builds, right?
blissful pollen
Yup I hadn't given it an approval yet waiting on that, and it has to go alongside the python busdevice library too. Both do look good now
blissful pollen
It's a change to the adafruit_busdevice module. So it is being changed both in the core and in the library that is used for the smaller builds
tulip sleet
did you cancel the bad build?
blissful pollen
Yes I did
blissful pollen
Thanks!
tulip sleet
i will keep an eye on the merge action to make sure it works: https://github.com/adafruit/circuitpython/actions
tulip sleet
@blissful pollen Merge actions almost done. Looks fine. The espressif builds were all fine.
blissful pollen
<@!228322543507865601> Merge actions almost done. Looks fine. The espressif buil...
Awesome. Thanks for helping on that
tulip sleet
np
mortal mica
Is there anything specific to do to get my board into the 7.0.0 release? I expected to see to the PyKey60 download page to have 7.0.0 available. Did the merge just missed the release by a couple of hours or is there something else I am missing?
analog bridge
Is REPL over hardware UART possible?
tulip sleet
Is `REPL` over hardware `UART` possible?
There are some compile flags to turn it on. It's only implemented for certain boards.
analog bridge
There are some compile flags to turn it on. It's only implemented for certain bo...
// in mpconfigboard.h
#define DEBUG_UART_TX
#define DEBUG_UART_RX
are you referring to this? ๐
I saw this in microbit_v2
tulip sleet
yes, I thought there was something else, but I looked and I can't find it.
tulip sleet
Is there anything specific to do to get my board into the 7.0.0 release? I expe...
Yes, it was merged after rc.3. We expect to do a 7.1.0 or something like that relatively soon, but there's no schedule. The "Absolute Newest" should be fine.
analog bridge
@tulip sleet I can't get REPL over UART with those defines on espressif.
tulip sleet
<@!329766224093249548> I can't get `REPL` over `UART` with those defines on espr...
I don't think it would be that hard to add. There's a CIRCUITPY_SERIAL_BLE, and the logic for UART would be similar. There was multiterminal but we don't use it for anything. It was for ESP8266, I think.
I added a CIRCUITPY_CONSOLE_UART but it doesn't do anything.
The name of that maybe should be changed, or the name of CIRCUITPY_SERIAL_BLE should be changed. I don't remember now.
analog bridge
thanks dan, I'll take a look at how CIRCUITPY_SERIAL_BLE is implemented, I need the repl over uart thing for esp32c3
tulip sleet
Scott may have a mental plan about this
lone sandal
crimson ferry
Does esp32s2 common-hal code need to feed the WDT, or yield in some way sooner than some interval?
analog bridge
@slender iron if I request 0x6d44_4333 id... then will the following be the entries in the board config?
CIRCUITPY_CREATOR_ID = 0x6d44
CIRCUITPY_CREATION_ID = 0x4333
jaunty juniper
I believe both creator and creation are 4 bytes long, if you request the creator ID 0x1337_BEEF, you use that as the creator ID, and whatever 32 bit value as the creation ID
slender iron
yup, what neradoc said. you just need to request the creator ID and then decide on your own creation id
analog bridge
okay so...
CIRCUITPY_CREATOR_ID = 0x6d444333
CIRCUITPY_CREATION_ID = any random 32 bit value
@idle owl were we doing something global about this f-string check? It's affecting some PR run: https://github.com/adafruit/Adafruit_CircuitPython_GPS/pull/67/checks?check_run_id=3702332800
tidal kiln
@onyx hinge yah. stuff is happening for that. kattni would know current state of progress though.
idle owl
<@!330227457296957440> were we doing something global about this f-string check?...
Yes. If the author is comfortable, they can do it in their PR. Is it otherwise ready to merge?
Might make more sense to simply merge it failing before we run the script, really.
onyx hinge
it was marked as being a new contributor PR so I would probably not ask for it
idle owl
Is it ready to merge now?
tidal kiln
if enabled, could push changes to their pr
idle owl
The script will be run today.
That's why I'm saying we can merge it now failing.
Or it may have conflicts once the script is run.
onyx hinge
I don't know enough about GPS, I think it's a safe-ish change though
tidal kiln
ah. got it. don't even worry about the change. it'll happen automagically later.
onyx hinge
that's fine then
idle owl
@tidal kiln @onyx hinge The script has been run, so all the libraries should have that update now.
onyx hinge
y'all are great, thanks!
tidal kiln
cool. thanks. the CI dragons have been put back to sleep...for now.
idle owl
For now, indeed.
tidal kiln
and on a friday ๐
idle owl
Welp. ๐คท๐ปโโ๏ธ It needed to be done.
idle owl
@gilded cradle Did the 2.9" Eink Displays guide get updated or is it new? The info on the bottom of the guide is inconsistent.
gilded cradle
@idle owl kind of in between. It's a new guide that drew from existing pages. Which info are you referring to?
idle owl
@gilded cradle The "This guide was first published on $date" and then the "This guide was last edited on $date" - the published date as 22 September, the last edited date was 16 September ๐คท๐ปโโ๏ธ
@onyx hinge As long as you're up for helping, I have code that mostly does what I need, except for one thing, and the only way I know how to deal with that is a simple state machine, and I think it's a terrible kludge that I keep perpetuating when I shouldn't be.
I need it to only send spacebar once, and not again until you've backed off and gotten closer again.
Right now, it repeatedly sends spacebar as if you're holding it down, which will make you fail the game.
gilded cradle
Oh yeah @idle owl, not sure why. Maybe because I used the copy feature from another guide.
idle owl
Oh yeah <@!330227457296957440>, not sure why. Maybe because I used the copy feat...
Ah ok.
onyx hinge
@idle owl always losing at the game isn't great ๐
idle owl
I'll have to send another PR to the newsletter to get it in there, I thought it was old. Thanks @gilded cradle
@onyx hinge Ok, that's fixed. So... how to only send it once.
Without using my "debounce" state machine code.
Here is what I'm referring to, written for the Circuit Playground library: py button_a_pressed = False while True: if cpb.button_a and not button_a_pressed: # If button A pressed... print("Button A pressed.") [...] # The code you want to execute on the button pressed event. button_a_pressed = True # Set to True. time.sleep(0.05) # Debounce. if not cpb.button_a and button_a_pressed: # On button release... button_a_pressed = False # Set to False. time.sleep(0.05) # Debounce.
Kludgy. ๐
Or is that acceptable?
I tried both press and send.
Neither one only sends once.
onyx hinge
it's probably fine?
idle owl
Hmm fair enough
onyx hinge
I'm surprised press sends again, or rather, that press again does the equivalent of release+press
(it's not what the source says to me, but if you tested it then anything I thought I knew is suspect)
idle owl
Hmm.
I'll try again.
It might be my code.
Making it keep sending.
Do you want to see the code? It's probably my code being wrong.
onyx hinge
if you want to share it in its current state
idle owl
"Want" is a strong word. ๐
initial_proximity = apds.proximity
while True:
print(apds.proximity)
current_proximity = apds.proximity
if current_proximity > initial_proximity and current_proximity > 127:
pixels.fill((255, 0, 0))
keyboard.press(Keycode.SPACE)
initial_proximity = current_proximity
elif current_proximity < initial_proximity and current_proximity < 127:
pixels.fill(0)
keyboard.release_all()
initial_proximity = current_proximity
time.sleep(0.2)```That doesn't include setup obvs.
That if is true more than it should be, I think.
onyx hinge
so your pixels are filled like you want, but the virtual keyboard isn't doing what you want?
idle owl
Well, the pixels aren't a good example because if they keep "filling", I wouldn't notice.
Because they're already on
onyx hinge
but it's red when the hand is close and unlit when the hand is removed
jaunty juniper
it doesn't look like it would release unprompted, so it should hold the space bar down
idle owl
They do turn on appropriately, but the space is initially sent appropriately too
idle owl
it doesn't look like it would release unprompted, so it should hold the space ba...
I don't think I quite follow.
Should it release inside the if ?
jaunty juniper
it should hold the space bar until you back away
idle owl
That's what it's doing. That's not what I want it to do.
So you're correct on what it's currently doing
onyx hinge
do you want it to act like it taps space just once when you first move your hand down?
idle owl
I think so?
If you hold space down playing this game, you die quickly ๐
So you need to only send it once when you want to
onyx hinge
then yes you'll need to use a state machine
or debounce or whatever you like to call it
jaunty juniper
and use send
onyx hinge
yes and use send
idle owl
Ah ok. I guess I always thought I was doing it wrong because I wrote that code really early on in my learning process.
And a lot of that code I wrote early on is... not great.
onyx hinge
there are a million different ways to phrase it but yours works
was_pressed = False
while True:
pressed = read sensor or whatever to get a True or False value
if pressed and not was_pressed:
do the actions for newly pressing a key
was_pressed = pressed
I tend to write something like that, personally
idle owl
Problem here is, the proximity sensor is not a True/False situation.
I guess I need to make it one.
onyx hinge
maybe pressed = current_proximity > initial_proximity and current_proximity > 127 ?
(I think there's something weird about the double comparison but that's a separate thing)
idle owl
OH!
jaunty juniper
it looks like you used current_proximity > initial_proximity as a way to do the debouncing, but it wouldn't work since it would still trigger when moving away (or in)
idle owl
Yes that.
I just figured that out.
I was trying to use that to make sure it only triggered when you were moving the right direction.
Maybe it doesn't matter.
jaunty juniper
I would remove that comparison
idle owl
Still sending repeated spaces. sigh.
jaunty juniper
(since it will always be true when the other condition becomes true, but also some other times)
or is the direction you move important for the game ?
analog bridge
Adafruit CircuitPython 7.0.0-78-gb8aa2b9cc-dirty on 2021-09-25; MicroDev microC3 with ESP32C3
Board ID:microdev_micro_c3
Auto-reload is on. Simply save files over USB to run them or enter REPL to disable.
code.py output:
Hello World!
Code done running.
Press any key to enter the REPL. Use CTRL-D to reload.
finally REPL on esp32c3 ๐ฅณ
idle owl
Congrats! Nicely done!
analog bridge
@tulip sleet follow-up from our previous conversation... REPL over hardware UART not working was espressif specific... common_hal_uart_never_reset() wasn't implemented properly.
I'll PR with a fix soon.
tulip sleet
<@!329766224093249548> follow-up from our previous conversation... `REPL` over h...
glad this was not generic, and glad you got it working. If you think the switching on/off should be cleaned up, that could be another thing to do, but you're probably on a particular path
onyx hinge
@idle owl I don't have a prox trinkey so I can't exactly follow along with you ... so instead I wrote a program using the encoder button of the macropad. ```python
import time
from adafruit_macropad import *
macropad = MacroPad()
old_pressed = False
while True:
pressed = macropad.encoder_switch
if pressed and not old_pressed:
macropad.keyboard_layout.write(' ')
old_pressed = pressed
time.sleep(.001) # I found that .2 was too long
what is the game btw ?
onyx hinge
@idle owl my goal is to try to separate the logic about "when to send a space" given a nice clear button press from "how to turn the distance sensor into a nice clear button press" and figure out which one is the problem
@jaunty juniper chrome://dino/ in google chrome / chromium browser
just enter that in the address bar to start the game
jaunty juniper
oh that's what the hiibot came pre-installed with
onyx hinge
yay!
onyx hinge
it is a cause for rejoicing!
idle owl
But it's not saving and reloading
I had to manually reload within the serial console to get it to update
onyx hinge
wait that's weird
onyx hinge
does it do it still now that it knows you're watching?
onyx hinge
are you on 7.0.0?
onyx hinge
does it persist after you use the reset button?
idle owl
Don't know, haven't tried yet
onyx hinge
we're getting v. close to "you get to file an issue now" ๐
onyx hinge
understood
idle owl
Now it's reloading.
Sigh.
Not worth filing an issue, imo. But wow that was annoying.
I'm not sure I've ever run into that before.
tulip sleet
i thought I saw something like this yesterday, and was really confused. But I could not reproduce it.
I think I was getting a USB error really early.
idle owl
I'm not great at this game, but I am worse with this input method. ๐
I need to ๐ฎ . BBIAB.
onyx hinge
@idle owl do try decreasing the sleep time, 1/5 of a second is an eternity in this game
jaunty juniper
by the way you can use a predicate with adafruit_debouncer, I think it would work:
distance = Debouncer(lambda: apds.proximity > 127)
while True:
distance.update()
if distance.rose:
keyboard.send(Keycode.SPACE)
pixels.fill((255, 0, 0))
if distance.fell:
pixels.fill(0)
Hmm. Fair enough.
This is easier for me to explain ๐
And actually, not sure debouncer would fit.
This one is tight.
jaunty juniper
oh right trinkey
idle owl
ya
jaunty juniper
too bad, I liked the idea of rose/fell having a double meaning ๐
idle owl
Agreed! And it would have worked! Except for me explaining lambda ๐
It's destined for a guide, as is probably obvious.
gilded cradle
Ah ok.
I realized what happened. I wrote that page on the 16th, but because I had to write drivers for the UC8151D eInk display, it didn't go live for almost a week and so the dates are correct.
idle owl
I realized what happened. I wrote that page on the 16th, but because I had to wr...
Ok! It's alright. It went into this week's newsletter, I don't think it was in last week's. So it all works out.
gilded cradle
Awesome
manic glacierBOT
manic glacierBOT
I checked the corrections against here, but why is there a D1, which is not on the silk nor the diagram?
onyx hinge
@tulip sleet report_ids = (0,) doesn't seem to be accepted
ornate breach
@analog bridge how stable is your ESP32-C3 port? Iโve got one of the Engineering samples from early this year and kind of want to try it out.
analog bridge
<@747673663419187281> how stable is your ESP32-C3 port? Iโve got one of the Engi...
I haven't done a lot of testing as currently only REPL interaction is possible over hardware UART but please feel free to give it a shot... most things should work.
One more thing, I am doing all the testing using the latest revision 3 of esp32c3, you might have to tweak the sdk config a little to specify the revision of your chip.
analog bridge
It appears when you interact with the chip using esptool.
ornate breach
I figured as much. Thanks for the work youโre doing in the C3 port
Iโve wanted to do more with the C3 but Iโve been waiting for the BLE port ๐
analog bridge
๐ ... ble is the hardest part.
analog bridge
I think following in board's sdkconfig should do
# CONFIG_ESP32C3_REV_MIN_0 is not set
# CONFIG_ESP32C3_REV_MIN_1 is not set
CONFIG_ESP32C3_REV_MIN_2=y
# CONFIG_ESP32C3_REV_MIN_3 is not set
CONFIG_ESP32C3_REV_MIN=2
cool, i'll pull down your repo and build it and try it out
analog bridge
use this branch instead of the PR one... I have got board files in this one
https://github.com/microDev1/circuitpython/tree/microdev-micro-c3
ornate breach
will i flash it the same way i would the bin for say an s2 using esptool?
analog bridge
yes, its the same... there were some changes with esptool config flags, they have been taken care of in the Makefile
ornate breach
okay, so just sh esptool.py --port /dev/cu.usbxxx --after=no_reset write_flash 0x0 firmware.bin
analog bridge
or you can use make BOARD=... flash
ornate breach
oh didn't know you could do that lol
analog bridge
I prefer the make one as it takes care of all the flags.
ornate breach
doing that, do i need to specify the com port target?
analog bridge
oh yeah, I forgot that... make BOARD=... PORT=... flash
ornate breach
ah perfect
analog bridge
the default is /dev/tty.SLAB_USBtoUART
solar whale
That post is 3 years old ๐
ornate breach
That post is 3 years old ๐
I know, just sad
I keep running into issues building
solar whale
I keep running into issues building
Which board are you unable to build?
ornate breach
Iโm trying to test out the ESP32-C3 stuff microDev is working on but I need to add the riscv tool chain and Iโm following the esp idf instructions for mac install but Iโm just running into build issues with crosstool-ng
solar whale
Hmm.. Iโve not had issues recently esp32s2 on my Mac. Itโs be a few weeks since I tried it and wonโt be able to check until tomorrow morning. Iโll try it then. I have not tried the C3 at all, but can you build the S2?
ornate breach
Yeah, as far as I can tell
analog bridge
@ornate breach I can provide you with .bin for REV-2.
solar whale
@ornate breach I see that the C3 may need a different set of tools from the S2 -- I can build the S2 OK on my Mac, but have not tried C3.
But if I run ./install.sh esp32c3 in the esp-idf folder, it does not complain.....I tried downloading pr 5394, but I don't see a C3 board in the boards folder to try building. I I must have missed something and since I don't have the hardware I think I'll leave this to you to sort out. Sorry I could not be of more help.
analog bridge
But if I run ./install.sh esp32c3 in the esp-idf folder, it does not complain......
@solar whale use this branch instead of the PR one... I have got board files in this one
https://github.com/microDev1/circuitpython/tree/microdev-micro-c3
solar whale
<@!343570430042308609> use this branch instead of the PR one... I have got board...
hmm -- that seemed to be building OK on my Mac, but failed near the end with /Users/jerryneedell/.espressif/tools/riscv32-esp-elf/esp-2021r1-8.4.0/riscv32-esp-elf/bin/../lib/gcc/riscv32-esp-elf/8.4.0/../../../../riscv32-esp-elf/bin/ld: build-microdev_micro_c3/supervisor/shared/serial.o: in function `serial_early_init': /Users/jerryneedell/projects/circuitpython_microdev/ports/espressif/../../supervisor/shared/serial.c:72: undefined reference to `common_hal_busio_uart_never_reset' collect2: error: ld returned 1 exit status make: *** [build-microdev_micro_c3/firmware.elf] Error 1
analog bridge
hmm -- that seemed to be building OK on my Mac, but failed near the end with ```...
a pull from adafruit:main should fix this
solar whale
how do I do that?
analog bridge
have you cloned my fork?
solar whale
I cloned your fork and checkout the microdev-micro-c3 branch
analog bridge
okay then following should do
git remote add adafruit https://github.com/adafruit/circuitpython.git
git fetch adafruit
git pull adafruit main
yay! that worked ```
1280672 bytes used, 161120 bytes free in flash firmware space out of 1441792 bytes (1408.0kB).
0 bytes used, 8192 bytes free in 'RTC Fast Memory' out of 8192 bytes (8.0kB).
16 bytes used, 8176 bytes free in 'RTC Slow Memory' out of 8192 bytes (8.0kB).
0 bytes used, 32768 bytes free in 'Internal SRAM 0' out of 32768 bytes (32.0kB).
0 bytes used, 294912 bytes free in 'Internal SRAM 1' out of 294912 bytes (288.0kB).
So it does appear to be possible on a Mac!
ornate breach
<@!728676724522352733> I see that the C3 may need a different set of tools from ...
Building the C3 board in circuitpython gave me a compiler/asm error for not being able to find the risc-v esp g++/gcc and asm compilers
Might be a linking issue, not sure ๐คทโโ๏ธ
solar whale
no idea -- mine did not complain....
lone axle
@idle owl I think the scripts to make issues for missing types across all libraries are all set and ready to go any time. Wanted to give a heads up here because whenever I run it there will be a big wave of them created and they @mention both of us now.
idle owl
<@!330227457296957440> I think the scripts to make issues for missing types acro...
I see that. ๐ Thanks for the heads up! RIP my inbox. Can you run it on a subset initially to verify that it's working properly before setting it free on the entire bundle? Or is that a huge pain to do?
Or did you do that already?
lone axle
I see that. ๐ Thanks for the heads up! RIP my inbox. Can you run it on a subset...
I have not done that yet. I can give it a try. I think it's possible by cloning a local copy of the bundle and removing all but a few of the submodules from the local copy.
idle owl
I have not done that yet. I can give it a try. I think it's possible by cloning ...
If it's too much of an issue, don't worry about it. I'm sure it's fine. It's something I always ask of Dylan when running entirely new scripts on the libs, so I figured I'd ask you as well.
lone axle
I'll try it out and if it starts looking too crazy I'll skip it.
idle owl
Perfect.
@tulip sleet You around for a question regarding guide feedback I received? It's regarding the Windows 7 CP drivers.
I think the answer is delete and move on, but I wanted to give it a fair go.
tulip sleet
AFK at the moment but will get back to you
idle owl
Ok thanks!
turbid radish
?serverinfo
digital shoreBOT
adafruit
ornate breach
31k ๐
idle owl
CircuitPython Weekly meeting happening in about 1 hour 45 minutes. Everyone is welcome to attend! If you're interested in participating either by attending or providing updates, please add your Hug Reports and Status Updates to the notes document before the meeting. Hope to chat with everyone soon! https://docs.google.com/document/d/1rS5p7xF_UUpumitnW2mV-B3IL7vB6e7V2HIAkk7Zo9E/edit# <@&356864093652516868>
tulip sleet
@idle owl Point me to the guide feedback and I'll take a look.
idle owl
<@!330227457296957440> Point me to the guide feedback and I'll take a look.
Oh! That's even easier. Hold on, getting link.
DM'd.
Like I said, pretty sure it's a delete and move on, but I figured I would at least check with you.
lone axle
Alright, I ran it on 4 libraries, one package and one single file module for each of helpers and drivers type libraries. Looking over the issues it created now. As long as everything looks good with them I'll run it on the rest of the list.
@idle owl the 4 that it created look good to me. I'm going to try to figure out if there is a relevant rate limit for creating the issues and maybe add some delays to the process if so. But should be running it on the larger list starting shortly.
idle owl
<@!330227457296957440> the 4 that it created look good to me. I'm going to try t...
Great! Thanks for running the test.
I'll ignore email for a bit here ๐
lone axle
okay, it's running now. If I'm understanding the rate limit endpoint correctly it's showing 0 used of all types so it seems maybe creating issues with the gh CLI tool does not use anything that is limited. It's taking about 3 seconds per library, and stops for manual intervention on ones where the adafruit repo was forked from a different user to ask me which user it should use. I'm keeping an eye on it and doing that as needed. Should be done in 5-10 minutes I think if the pace holds.
idle owl
okay, it's running now. If I'm understanding the rate limit endpoint correctly i...
Ah fair enough. Sounds good.
lone axle
Ah, nope it started failing with GraphQL error: was submitted too quickly
lone axle
I will save the output from what it's done so far and start up the remaining ones in a little while.
solar whale
Yikes -- maybe I should do a "reply-all" to each of those e-mails from @lone axle ๐
lone axle
I think the permissions for this repo might be different from the majority of them: https://github.com/adafruit/Adafruit_CircuitPython_PCF8563 creating issue failed with GraphQL error: You don't have permission to add labels in this repository.
idle owl
Thanks. I'll take a look.
@lone axle Fixed. The CPLibians team wasn't there.
The team is there now.
Oi. Wonder how long that's been the case - Adabot doesn't work without that.
lone axle
I'll circle back and run it on that one at the end.
idle owl
Thank you!
lone axle
@idle owl another one with the permission error here: https://github.com/adafruit/Adafruit_CircuitPython_Ducky
idle owl
That's a recent one. Need to remember to mention it to Dylan too.
idle owl
<@!330227457296957440> another one with the permission error here: https://githu...
Fixed!
lone axle
Fixed!
another newer one with the same issue: https://github.com/adafruit/Adafruit_CircuitPython_Pixelbuf
idle owl
another newer one with the same issue: https://github.com/adafruit/Adafruit_Circ...
Fixed!
lone axle
Alright. We've made it through the whole list now. There are a few that it skipped or didn't render properly. I'll go through and get them all squared away later on this afternoon after the meeting.
idle owl
Alright. We've made it through the whole list now. There are a few that it skipp...
Excellent!
Internal meeting just ended. BRB!
idle owl
Hmm. Audio not right.
onyx hinge
uh oh
lone axle
@onyx hinge Yep. The main benefit that I use to checkout a PR branch:
gh pr checkout [PR #]
But I've also learned recently that you can leave reviews as well
gh pr review --approve -b "Tested successfully on PyPortal. Looks good to me"
It finds the active branch to know which PR to make it on.
onyx hinge
wow neat!
slender iron
I won't fight you for it ๐
lone axle
As I understand it, the difference is the community bundle is meant for libraries supported by individual's whereas the circuitpython org bundle is for libraries supported by multiple folks.
jaunty juniper
it's "Circuitpython Organisation" Bundle not ".org" as in github organisation, aka https://github.com/circuitpython
the name is confusing because of "org" vs ".org"
lone axle
One thing we do still need in the CircuitPython organization bundle is to get adabot (or some other tool) making the automated releases whenever one of the submodules has a new version. Right now the releases for CircuitPython organisation bundle are manual.
onyx hinge
@idle owl I wonder if Jerry should go before Melissa since the topics are related
but we need the intro first so everyone's on the same page
okay
@gilded cradle do you have a link to show the sorts of issues you're concerned about? what do the errors look like, etc.
gilded cradle
Let me check... It was a little while ago
heady island
(I'm just a user, is there a link to when these "calls" happen? It's very interesting)
ornate breach
You need to be part of the <@&356864093652516868> I believe to join
Sorry for the ping ๐
cobalt grail
Ping.
slender iron
they are mondays at 11am pacific 2 pm eastern
onyx hinge
@heady island we meet most mondays at 2PM Eastern (US) time, you can ask to join @ circuitpythonistas to get notifications or to speak in the meeting, or check out our online calendar https://github.com/adafruit/adafruit-circuitpython-weekly-meeting
@solar whale it adds about 30 bytes to an mpy file
slender iron
past meetings are available on youtube too: https://www.youtube.com/playlist?list=PLjF7R1fz_OOUvw7tMv45xjWp0ht8yNgg0
onyx hinge
irrespective of the quantity of type annotations, those are removed
cunning crypt
Sorry for the ping ๐
You can put the group name in backticks so it doesn't ping people.
heady island
thanks ๐
onyx hinge
If a function says def f() -> int: ... the -> int part is just ignored by circuitpython, so if it refers to typing.List then it doesn't matter that this doesn't exist
gilded cradle
@onyx hinge it was back and April and it looks like the GitHub Actions results have been discarded since then.
Here's the PR though: https://github.com/adafruit/Adafruit_Blinka_Displayio/pull/58
onyx hinge
jerryn said "if you add a comment, you need to run the regression test" ๐คฃ
solar whale
"should" ๐
onyx hinge
we'll give you a "should"er to cry on
lone axle
When I complete the annotations for a library this week I'll test on Blinka and microcontroller various scenarios to see if I can find out what if any possible issues may arise.
heady island
as a new user - a few blogs pointing me to this discord mentioned the PIO channel
If there was a msg to go to "new channel" it would be fine
slender iron
@heady island would you like me to add you to circuitpythonistas?
gilded cradle
Thanks
heady island
@slender iron sure! ty
lone axle
Thanks all, hope everyone has a nice week! ๐
idle owl
@slender iron The audio is borked on the video. My audio is fine, everyone else is consistently intermittent. Sigh. OBS did not indicate there were any problems.
onyx hinge
argh
idle owl
@slender iron Do I post it with messed up audio (it's pretty bad to listen to), or do we skip this week?
idle wharf
@idle owl I'll still listen to it ... but happy to test it and tell you if its unlistenable
idle owl
<@!330227457296957440> I'll still listen to it ... but happy to test it and tel...
Sure, thanks.
slender iron
I think post it still with the first sentence of the description acknowledging the problem
idle wharf
They can at least listen to Kattni's excellent intro audio and bail out when the others start talking.
The others are inaudible \ unable to be understood.
idle owl
Fair enough. I'll make it public and submit another PR to the newsletter.
lone axle
It turns out that SimpleMath and ProgressBar libraries both already have type information for everything in them. So those are our first two completed libraries ๐ . I'll do some testing on Blinka and a Microcontroller this week with these two in addition to another library that I'll fill in the types for, maybe a display driver to validate Blinka_Displayio environment.
idle owl
It turns out that SimpleMath and ProgressBar libraries both already have type in...
That's perfect. Thank you so much for looking into this and getting all of the issues made.
timber pivot
Not sure if this is the right channel but you all know the audio on CircuitPython Weekly for September 27th, 2021 video on YouTube is messed up, right?
lone axle
Not sure if this is the right channel but you all know the audio on CircuitPytho...
This is the right channel, and yep. There was an issue with the recording this week. There is a discussion up a little ways in this channel.
There are text notes available here: https://github.com/adafruit/adafruit-circuitpython-weekly-meeting/blob/main/2021/2021-09-27.md if you're interested in seeing everything that was discussed on the audio.
timber pivot
Thanks
lone axle
I will try to start making backup recordings again starting with next weeks. I was recording them for a little while but I got tripped up one week by having an extra discord window open and fell out of the habit after that.
idle owl
Not sure if this is the right channel but you all know the audio on CircuitPytho...
There's a note on the video, and in the notes document regarding the audio. Apologies. The recording software did not indicate any issues, so there was no way to know until the recording had been completed.
tidal kiln
@idle owl hey. any idea why no CP lib for this?
https://www.adafruit.com/product/1746
guessing because disco?
idle owl
<@!330227457296957440> hey. any idea why no CP lib for this?
<https://www.adafru...
No idea! Probably? No idea when it was disco'd. Might have been pre-CP.
I never got one, looks like.
tidal kiln
is opening an issue in bundle repo still how to request?
idle owl
Um.... hmm. I thought it was opening an issue on the core. Let me take a look. Hold on.
@tidal kiln Open an issue on the core and label it "library".
That's where we have other driver requests.
tidal kiln
ok. thanks. will do.
good to have to at least document, even if decide not to write
i'm not seeing any old issues requesting it?
core or bundle
orchid basinBOT
manic glacierBOT
lone axle
In situations where types come from existing modules within circuitpython but those modules are not used for anything other than typing would it make sense to do imports like this:
try:
from typing import List
import digitalio
import busio
except ImportError:
pass
presumably digitalio and busio would not raise an error. But importing them could take up memory right? So doing the typing import first and catching the exception will prevent digitalio and busio from being imported when the program runs on a microcontroller?
tulip sleet
In situations where types come from existing modules within circuitpython but th...
That makes sense but I would add a comment like # Used only for typing. before the imports to make this clear.
lone axle
Thank you
analog bridge
Is there any way to prevent lookup for a uf2 firmware file in CI for non-uf2 mcu's like esp32c3? Currently CI fails with the following error:
Cannot find file ../ports/espressif/build-microdev_micro_c3/firmware.uf2
Error: Process completed with exit code 1.
I'm not sure if it's there or if the information is duplicated somewhere else
analog bridge
I added the board in extension_by_board... doing a test now.
analog bridge
That worked! Thanks @jaunty juniper.
crimson ferry
I seem to have forgotten (or broken) something... after having an esp-idf binary on an esp32s2, I'm unable to flash CircuitPython. Anything I try just gets the BOOT drive back on reset. Is there a canonical set of esptool parameters in use these days?
solar whale
You may want to try erasing the flash before loading CP
crimson ferry
yes, thanks, that did it
solar whale
Great -- I've run into that a few times.
crimson ferry
I'm guessing it would have also worked if I had installed TinyUF2
crimson ferry
does anyone know what's up with current main esp32s2 build errors related to undefined reference to 'port_i2s_reset_instance'? ...something to do with paralleldisplay
supple gale
hmm, not here. just did a compile with no errors .
crimson ferry
maybe due to CIRCUITPY_AUDIOBUSIO_I2SOUT ?= 0 to make space for DEBUG=1? I'll try a new scheme for space
supple gale
yes, that makes sense
crimson ferry
when there's a module dependency like that, is it typical for turning off a module to cause build errors, or is it more manual to keep track of those things? (I don't see a new flag for parallel display)
hm, re-enabling CIRCUITPY_AUDIOBUSIO_I2SOUT didn't do it, I'll just comment out the three lines of relevant calls for now
supple gale
I think it needs CIRCUITPY_AUDIOBUSIO
crimson ferry
ah, there they are, thanks!
supple gale
but not sure it should be called in any case. i leave that to someone who better understands what they have done with I2S
crimson ferry
I don't think they are called, but build gets unhappy
supple gale
its called in the deinit i think.
crimson ferry
oh, like from a hard reset?
ugh, so I infer that it's manual to know which modules depend on others when turning off modules (for space)
supple gale
i guess, freeing up the i2s instance that was created for parallelbus.
crimson ferry
thanks for your help, back in business ๐
supple gale
๐
lone axle
is it possible to see the resulting html files from the sphinx docs build from a PR branch before the PR is merged? or is that only visible once it gets merged and is then shown on the readthedocs pages?
jaunty juniper
๐ค I would build them locally
lone axle
๐ค I would build them locally
The thing is, I think I'm seeing different results locally compared to what is getting generated by actions. I'm trying to determine specifically what the ones generated by actions will look like because I suspect that mine locally are different and not correct.
Now I'm seeing different results than I was getting the other day even though I don't believe anything int he environment or code should be different ๐
jaunty juniper
oof
onyx hinge
@lone axle I don't think the doc artifacts from github builds go anywhere, but that doesn't really matter as what is shown on readthedocs doesn't come from there either
.. it comes from readthedocs own process
lone axle
I was having this strange difference between Sphinx running in actions vs. running locally. Actions completed successfully, but locally failed.
The user that made that PR also saw the same error locally that I was seeing.
But actions completed successfully. And now if I go back and run the build on the same code locally it's succeeding. I don't know what could have changed ๐
onyx hinge
https://discord.com/channels/327254708534116352/537365702651150357/8917394881942...
That looks like a problem with a mocked module. If that doesn't mean much to you I can say more..
lone axle
That looks like a problem with a mocked module. If that doesn't mean much to you...
I agree, and the error does make sense to me. And I was able to get passed one of the errors by adding microcontroller to the mock list. But it's not there in the actual repo. So I don't understand how it didn't have the same error when it ran inside of actions.
onyx hinge
the fake version of a module provided in this way works in some respects, but not all. So for the error you linked to,
"/home/timc/repos/circuitpython/Adafruit_CircuitPython_BLE_Radio/adafruit_ble_radio.py", line 51, in _RadioAdvertisement
match_prefixes = (struct.pack("<BH", 0xFF, _ADAFRUIT_COMPANY_ID),)
I am guessing that _ADAFRUIT_COMPANY_ID is a mocked object, and that simply can't stand in for the kind of object that struct.pack requires.
lone axle
it's not listed in conf.py though: https://github.com/adafruit/Adafruit_CircuitPython_BLE_Radio/blob/2bd9f0ddc2206a3e1c16ed383a4e0f68915e1d8a/docs/conf.py#L28 is there somewhere else that those can come from when it runs in actions?
onyx hinge
from micropython import const ... _ADAFRUIT_COMPANY_ID = const(0x0822)
so if from micropython import const fails, then const is a mocked function, and its return value is a mocked object. That mocked object, whatever it is, doesn't work with struct.pack. from micropython import const should come from adafruit blinka, I suppose?
lone axle
it did seem related to const() to me. I noticed that when I was getting that error locally if I removed the const() call wrapping the value that it was not having the error any more.
onyx hinge
requirements.txt lists Adafruit-Blinka so from micropython import const seems like it should work
lone axle
I do think blinka contains a placeholder for that. I believe that my local environment had the same version of Blinka installed as the actions did though. But I'm also not getting the error now when I run it on the same library
hmm maybe I was wrong about it being installed at the time I guess. If I uninstall it now I can force the same errors to appear
okay. It specifically needs to have adafruit-blinka installed and not to have microcontroller in the mock list
onyx hinge
what's odd to me is, conf.py doesn't specify to mock the micropython module, so it should have made the import an error instead. Or maybe a mocked module is mocked as a first resort instead of a last resort.
the docs are not super clear on this point
lone axle
Maybe I started without blinka installed, added microcontroller to the mock list. Then installed blinka but at that point still ended up with argument is not an integer I needed to go back and remove microcontroller from the mock list.
It's probably an entirely different issue. But this is the other thing that came up tonight that made me believe that the actions sphinx run can result in different outcomes / files than the local one. For this display driver the current file in readthedocs shows the correct arguments for the class:
But when I run the sphinx build locally against main I end up without the bus argument:
Maybe the deal is that readthedocs process differs slightly from Sphinx.
onyx hinge
do you have adafruit-blinka-displayio installed?
I ask because the only thing that "smells like a mock" is the fact that ILI9341 subclasses displayio.Display
lone axle
I do
I wonder if this configuration https://github.com/adafruit/Adafruit_CircuitPython_ILI9341/blob/9a3e9c3adeb5949567271cd99de349a08db9e47b/.readthedocs.yml#L6 is causing the docs to get built with python 3.0 inside of readthedocs.
I'm not sure if I'm understanding the documentation for that correctly, but I see something are listing a specific version like 3.5 or 3.7
maybe it installs python3 and ends up with whichever version happens to be marked as stable inside of it's VM which I guess would depend on the specific OS it's using.
Anyway I appreciate you taking a look and offering up some ideas. With your help I think I've understood what happened to me the other day with the BLE_Radio library, which was definitely more concerning to me since it was error vs. no error. This difference is relatively minor in comparison.
onyx hinge
Locally, I have to remove displayio from autodoc_mock_imports to get the correct signature.
lone axle
Ah! nice find. same here.
but in actions it would fail since it wouldn't have adafruit-blinka-displayio maybe
onyx hinge
it's listed in requirements.txt so it should
lone axle
oh, actually that is in requirements, yep
It makes me wonder how the copy that is currently in readthedocs came out correctly. But I can probably put that curiosity out of mind. I have a PR for typing open in that driver, I'll try removing the mock for displayio from conf and see if it remains happy.
onyx hinge
readthedocs says projects created a long time ago use an old sphinx version
Downloading Sphinx-1.8.5-py2.py3-none-any.whl (3.1 MB)``` and in factwhich could easily make its behavior different than what you see locally.
lone axle
Oh wow, I don't know what their release pace is like but that does seem quite old. That is very likely it
lone axle
I wonder if projects created before that can specify to use a newer one.
onyx hinge
and we're currently at 4.2.0
I have to assume so, but I don't seem to have the privileges to look or change it for this particular repo
lone axle
https://github.com/readthedocs/readthedocs.org/issues/6296 maybe including sphinx with a version in requirements.txt
You have solved so many mysteries for me, I have suspected differences on a few other occasions but never chased down the specifics this far. Thank you so much for taking the time! I really appreciate it.
onyx hinge
I have a RTD project registered in 2017 but I don't see a setting for sphinx version in the admin pages
okay, goodnight
it sounds like you should be able to put an explicit sphinx requrement and have it honored, yeah
manic glacierBOT
analog bridge
@teal condor you can get more help in #help-with-circuitpython channel... this channel is primarily used for circuitpython development work. ๐
teal condor
Thanks! I'll ask my question there itself, won't clutter here ๐
manic glacierBOT
tulip sleet
@idle owl I wrote a library https://github.com/adafruit/Adafruit_CircuitPython_Radial_Controller which I expect to use in a Learn Guide, showing the code. But maybe I should just put the code in the Learn repor? Do we have guides that point to code in libraries, not just in the Learn repo? I have a sample boot.py and a sample code.py to show, but they are in examples/ in the library (and are named differently.
lone axle
<@!330227457296957440> I wrote a library <https://github.com/adafruit/Adafruit_C...
There are guides that point to library repos instead of specifically the Learn System repo. Most of the "Primary Guides" for specific hardware devices that I've run across embed example code from the driver library repo instead of the Learn System repo. For example, this page: https://learn.adafruit.com/adafruit-128x64-oled-featherwing/circuitpython embeds from: https://github.com/adafruit/Adafruit_CircuitPython_DisplayIO_SH1107/blob/main/examples/displayio_sh1107_simpletest.py
tulip sleet
There are guides that point to library repos instead of specifically the Learn S...
Thanks! I was looking for an existence proof, and that is exactly what I need.
onyx hinge
@tulip sleet and once the library is in the bundle, within 24h you should have a folder-image for the examples at https://adafruit.github.io/Adafruit_CircuitPython_Bundle/
I think it happens each time the bundle is released
idle owl
<@!330227457296957440> I wrote a library <https://github.com/adafruit/Adafruit_C...
Sounds like you're good to go?
tulip sleet
Yes, thanks!
idle owl
Great.
idle owl
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/1_ZwlimMfNiDPSwWszVzGFMKQzGXSjmb5TzzhQ97kiOY/edit
idle owl
@tulip sleet Are you going to submit your lib to the bundle?
tulip sleet
yes, i'm just trying to get readthedocs set up. it's not showing up as a subproject to add yet
idle owl
yes, i'm just trying to get readthedocs set up. it's not showing up as a subproj...
Did you add adabot as a maintainer? So I can take a look.
tulip sleet
not yet, but I was looking at the HID library as an example, and it doesn't have adabot as a maintainer either, which is weird
idle owl
Yes it does?
idle owl
No, on Read the Docs itself.
Is where you add maintainers to docs.
Did you set up your lib on RTD?
(brb)
tulip sleet
oh, yes, I was looking in Subprojects by mistake
idle owl
Back.
tulip sleet
ok, i added the project, added the webhook, and added it as a subproject
idle owl
Excellent. Did you add adabot as a maintainer on RTD?
tulip sleet
I'm logged in as adabot, not me, I always do that
tulip sleet
it's one less step
idle owl
Yep.
tulip sleet
do i need a discord webhook too?
idle owl
No.
tulip sleet
I think I'm all set then, I'll make a PR to add it to the bundle. A little weird the code is not reviewed, but that's always true when I create a new lib
idle owl
I think I'm all set then, I'll make a PR to add it to the bundle. A little weird...
We'll find out if there's bugs once folks are using it ๐
tulip sleet
yeah ๐
idle owl
Complaint driven development.
tulip sleet
lol
idle owl
@tulip sleet Merged your bundle PR. Thanks for submitting it!
tulip sleet
thanks! onward
tidal kiln
anyone know if it's possible to restart a CI run that's like 15 months old and the logs are gone?
idle owl
anyone know if it's possible to restart a CI run that's like 15 months old and t...
Link?
tulip sleet
it will still run against the old commit it's trying to merge to, you'd have to push a new commit to get it try against the latest
idle owl
Yeah that would work. Push an empty commit should force it.
tidal kiln
not my pr though. don't worry about this if not quick/easy answer. it's not a CP thing.
would like to see all the output from the CI run on their last commit
tulip sleet
if they allow maintainer pushes, you could clone their repo and still push an empty commit. or just edit in GitHub and add a period to the end of sentence or something like that
tidal kiln
oh yah. that trick. i'll try that. thanks.
idle owl
not my pr though. don't worry about this if not quick/easy answer. it's not a CP...
Yeah looks like simply doing the restart from GitHub won't work. Edit a file would work.
tidal kiln
dan's trick worked. thanks!
yah, looked all around for a "restart" option. did not find.
idle owl
Same.
tidal kiln
thanks for looking!
idle owl
You're welcome!
idle owl
Sigh. I have a Feather M4 that isn't showing CIRCUITPY.
Unplug and replug fixed it. Ugh.
idle owl
@tidal kiln OK, I'm tagging you in. How do I pull a GPIO pin low in CircuitPython? e.g. simulate it being tied to gnd. (This is for that encoder again.)
I think I did it right. But the encoder isn't being read.
The buttons work, but the rotary encoder is doing nothing.
Double checked the pins.
Similar code worked for JP. Except he tied the pins to GND directly.
And my center button is intermittent which is even weirder to me.
Maybe this hack isn't possible on CP. Arduino sets the pins to output, pulled low. Can't do output and low in CP.
I'll check in again in the morning.
manic glacierBOT
manic glacierBOT
tidal kiln
@idle owl guessing its something else, and probably something simple? should be doable. just your basic
pin.direction = digitalio.Direction.OUTPUT
pin.value = False
@tulip sleet hey. there was a "CIRCUITPY not showing up on win 10" thing in #help-with-circuitpython last night. i think maybe it ended up being a known scenario? looks like they got it working eventually.
tulip sleet
<@!329766224093249548> hey. there was a "CIRCUITPY not showing up on win 10" thi...
Could have been a badly corrupted CIRCUITPY, or Windows seeing too many errors on the USB port and then refusing to see the device. It does look like Uwe Sieber's Device Cleanup fixed it
tidal kiln
yah. probably the latter.
manic glacierBOT
Hello all,
I am having nearly 10 - SCD_30 Sensors for measuring CO2 Value, Temperature, and Humidity for an project.
I was recording succesfully all the values without any problems making simple circuit with rapsberrypi:
Vcc,grnd, SDA, SCL and SEL.
Suddenly i am getting value of 0.0 for CO2 alone. i am not sure why its happening. But works fine with different sensor with ...
tidal kiln
@tulip sleet sry for asking yet again, but what is latest approach for flash erase? (when REPL access not possible) still here?
https://learn.adafruit.com/welcome-to-circuitpython/troubleshooting#old-way-for-the-circuit-playground-express-feather-m0-express-and-metro-m0-express-2978458-30
tulip sleet
<@!329766224093249548> sry for asking yet again, but what is latest approach for...
That is the only way if REPL access is not possible. You might be able to enter safe mode by the "slow" double-click method. So, "old" was only with respect to reformatting a corrupted filesystem. It's still the only way when there's absolutely no way to get to the REPL.
tidal kiln
oh. yah. safe mode.
for the other approach, if there isn't a UF2 eraser for target board, then what?
im 0 for 10 trying to get into safemode
stuck elbow
I usually only get there by mistake
tulip sleet
im 0 for 10 trying to get into safemode
In CPy 7, click reset, and then click again in about half a second, while the RGB is yellow. You can't wait for the yellow, our reaction times are too slow.
The other way is to build one of my CIRCUITPY-ERASER-DANGER uf2's, which involves changing one boolean. I can make you one or tell you how.
tidal kiln
if they are starting with an older firmware version, could they update ok? they still won't get CIRCUITPY, but firmware will update to 7? (to make getting into safe mode easier)
board is qt py m0
tulip sleet
sure
tidal kiln
Running in safe mode! Not running saved code.
You are in safe mode because:
You pressed the reset button during boot. Press again to exit safe mode.
Press any key to enter the REPL. Use CTRL-D to reload.
Adafruit CircuitPython 7.0.0 on 2021-09-20; Adafruit QT Py M0 with samd21e18
>>>
ok, yah, that was easier with 7
tulip sleet
we lengthened it from 700ms to 1sec
To create a CircuitPython that reformats the filesystem every time it starts, in circuitpython/main.c, change the false to true and rebuild:
// Create a new filesystem only if we're not in a safe mode.
// A power brownout here could make it appear as if there's
// no SPI flash filesystem, and we might erase the existing one.
filesystem_init(safe_mode == NO_SAFE_MODE, false);
any chance that could get added to the ci so one gets built for all boards?
tulip sleet
we had a long discussion about this: https://github.com/adafruit/circuitpython/pull/4722 and ultimately didn't do it
jaunty juniper
make clean doesn't delete "index.rst" files, does it ? is there a way to remove them after a make html ?
lime trellis
@tulip sleet I'm hitting the HardFault_Handler when developing samd fake deep sleep. But only on some boards. What's the easiest way to debug this? Do I need to dust off GDB?
tulip sleet
<@!329766224093249548> I'm hitting the `HardFault_Handler` when developing samd ...
That is the easiest way. Set a breakpoint on HardFault_Handler or, better reset_into_safe_mode(), and you'll see what led to the hard fault.
tulip sleet
`make clean` doesn't delete "index.rst" files, does it ? is there a way to remov...
yah, it really should. I think right now they are .gitignore'd.
jaunty juniper
yep, specifically /shared-bindings/*/**/*.rst is ignored
tulip sleet
@onyx hinge The generated screenshot images for project bundles: where are they committed? I see the form of the github.io URL, but I can't actually find where they are pushed to, despite reading images.yml in the Bundle repo.
onyx hinge
@tulip sleet the github.io page is showing https://github.com/adafruit/Adafruit_CircuitPython_Bundle/tree/folder-images
tulip sleet
OH, a separate branch
onyx hinge
if git commit -m"update images"; then git push -f origin HEAD:folder-images; fi the name of the branch is hidden in this git command where nobody will ever find it
tulip sleet
yes, I never found it ๐
onyx hinge
github pages likes to serve everything from a particular branch of a repo, usually not the main one
tulip sleet
Maybe I will add that to the README, not sure it's documented anywhere
onyx hinge
agreed, it needs to be documented. and maybe rolled out to the other bundles.
tulip sleet
thanks!
onyx hinge
you're welcome
blissful pollen
Has anyone had any trouble looping an MP3 in 7.0? I pulled out an old project (HallowWing M4) and even with loop=True it stops playing after one loop
tulip sleet
Has anyone had any trouble looping an MP3 in 7.0? I pulled out an old project (H...
It's possible we missed testing that. Could you file an issue with a simple example? Thanks.
blissful pollen
It's possible we missed testing that. Could you file an issue with a simple exam...
Will do, just wanted to check first, haven't had much of a chance to look into it further
idle owl
<@!330227457296957440> guessing its something else, and probably something simpl...
๐ That was it!!!!! Thank you so much. Went around in so many circles with that. The solution is so obvious but was nowhere on my radar.
Now to figure out how to make the pixel move with the encoder.......
tidal kiln
cool. np.
ping me if you want if you run into other oddness
is that the only wiring connection you plan to show in the guide?
idle owl
That was the plan yeah. Limor said it was fine when I checked in to verify.
I guess I could show it "wired up" typically as well. But I'll have to test it. <grump>
tidal kiln
i'd suggest mentioning the general nature of the hack, like "doing this just so it can be plopped down on a bread board next to a feather"
tidal kiln
otherwise it will be interpreted as the required way to wire it up
idle owl
Valid point.
tidal kiln
put in the text. put it in a call out box. make it bold. make it blink. make it scroll.
idle owl
Call show().
@tidal kiln Don't suppose you want to help me with the pixel code? I'm trying to port an Arduino sketch and I'm not quite getting it. I have it lighting up a pixel as it moves around, but either it's not turning off the other pixels after it lights them up, or it turns them off but only flashes the pixel as it moves.
tidal kiln
sure. same ard code from yesterday?
idle owl
Yes.
Here's what I have in CP so far: py while True: position = encoder.position if last_position is None or position != last_position: print("Position: {}".format(position)) last_position = position pixels[position % len(pixels)] = (0, 150, 0) pixels.show()
This lights them up but does not turn them off behind it.
If I add an else: pixels.fill(0) after that, it only flashes it as it moves.
(with show())
tidal kiln
is position generally working? like if you just print that value and spin dial - changes as expected?
idle owl
Yep, it is now, thank you for help with that.
tidal kiln
ah. just add a pixels.fill(0) in there.
pixels.fill(0)
pixels[position % len(pixels)] = (0, 150, 0)
pixels.show()
Now nothing lights up.
Oh
hold on
Magic!
I was so close!
Thank you ๐
I think I can figure out the rest. I'll ping you if I run into another wall.
tidal kiln
that's the equiv of the pixels.clear(); line in the ard sketch
idle owl
Ohhhhhh
idle owl
Tricksy
tidal kiln
and the arduino sketch is only using that conditional for the serial prints
tidal kiln
if (curr_rotary != last_rotary) {
Serial.print("Encoder value: ");
Serial.print(curr_rotary);
Serial.print(" direction: ");
Serial.println((int)direction);
}
everything else happens outside the conditional in the loop
idle owl
I'm not sure I know what you mean by conditional.
tidal kiln
your check here if last_position is None or position != last_position:
the pixel updating can be outside that
idle owl
Not sure I need auto_write=False to begin with.
Which would clean up the show()s.
I took it out for the encoder pixel, and it seems to be working ok?
tidal kiln
get position and button info
clear all the pixels
set pixel(s) based on position/buttons
show
repeat
i would. i'd turn auto_write off and call show at the end.
idle owl
Wait, what?
tidal kiln
you do several pixels sets - for the dial and then for each button?
idle owl
Yeah. But the button code is different. Because I created them in a list.
for i in range(len(buttons)):
if not buttons[i].value:
if i == 0:
print("enter")
pixels.fill((100, 100, 100))
# pixels.show()
if i == 1:
print("up")
pixels[0] = (150, 0, 0)
# pixels.show()
if i == 2:
print("left")
if i == 3:
print("down")
if i == 4:
print("right")
else:
pixels.fill(0)```Haven't set them all yet
And was trying with show() at the end
it does not function as expected.
With that code, with a show() after it at the same indent as the for, it simply turns off the LEDs when I press a button.
tidal kiln
did you enable pull ups on the inputs?
idle owl
Yep
button_pins = (SW1, SW2, SW3, SW4, SW5)
buttons = []
for button_pin in button_pins:
pin = digitalio.DigitalInOut(button_pin)
pin.switch_to_input(digitalio.Pull.UP)
buttons.append(pin)```They work. It prints the right thing when I press a button.
lime trellis
That is the easiest way. Set a breakpoint on `HardFault_Handler` or, better `res...
So I crash into HardFault_Handler on a pycubed board on builds WITHOUT DEBUG=1 but it runs fine on builds with DEBUG=1. Have you encountered this?
tidal kiln
@idle owl wanna move to #help-with-circuitpython ?
idle owl
Sure. Or we could move to a thread. Up to you.
#help-with-circuitpython doesn't seem busy. We can do that.
tidal kiln
hmm. let's try thread. need more experience using those.
tulip sleet
So I crash into `HardFault_Handler` on a pycubed board on builds WITHOUT `DEBUG=...
Oh, sure, many times ๐ . There are two common reasons I've encountered for DEBUG/no-DEBUG crash differences. One is that the bug is a second-order bug: something was smashed a long time ago, so reordering code, storage, etc., changes the bug. The other is that something that should be volatile is not volatile in the no-DEBUG build, but is accidentally volatile in the DEBUG build due to optimization differences.
You could try doing a -ggdb3 build with the regular optimization; it may already be set to do that. Can you get a backtrace? You can get a less-informative backtrace without debug symbols in the no-DEBUG case.
lime trellis
Oh, sure, many times ๐ . There are two common reasons I've encountered for DEBU...
ooo I'm going to look into possibly missing a volatile. In the meantime, how do I snag a backtrace from a no-DEBUG build? I'm using AS7 (which just learned they renamed to microchip studio haha)
tulip sleet
i have no idea how to do it in AS7. I always use gdb.
Maybe you can still set a breakpoint.
lime trellis
Maybe you can still set a breakpoint.
tried but it couldn't find the function. I'll thrash a bit and let you know
idle owl
@tidal kiln Bold and info-y enough?
tidal kiln
yah. hopefully.
maybe slightly different text
the more "normal" way would be tie COMA/B to GND
idle owl
Updated.
tidal kiln
The wiring shown here ties COMA/B to digital outputs set LOW instead of GND to allow for the direct-to-Feather connection.
?
idle owl
Hmm yeah.
stuck elbow
I wonder if it would be good to add a warning that this trick is not good for devices that take more power, because the gpio pins can only sink so much current.
idle owl
I don't follow. What devices that take more power?
onyx hinge
like, you couldn't use this trick for the GND of a neopixel strip
idle owl
I mean, it's going in a guide for this specific device.
We show wiring diagrams in guides for other devices.
So I don't think it's necessary to extrapolate here.
I guess if it comes up, we deal with it. But I think it's fine here.
stuck elbow
I think it would still be nice to explain when it's safe to repeat such technique, people are going to pick up this trick and try to use it in their own devices
I think this is actually a nice learning opportunity here, since there is a neopixel strip being used in the same project, so it could be mentioned that its gnd has to go to true gnd
idle owl
Hmm fair enough.
idle owl
@tidal kiln Tagged you on this as well as Limor, since you seemed familiar enough with it. https://github.com/adafruit/ci-arduino/pull/107
@tidal kiln Do you know how long it takes for updates to that file to propagate? Is it instant because it pulls from the latest version? Or is there latency on that. (Submitting code to Learn soon that will need this update.)
tidal kiln
unless it's cached, i'd expect instant
idle owl
OK that's what I was thinking too.
manic glacierBOT
crimson ferry
the ring buffer is hurting my brain... each element is limited to two bytes?
tulip sleet
yes, and the two-byte is hacked on. We got it from MicroPython
it could use a rework, but it's small
crimson ferry
do we have a queue-like structure that can handle arbitrary number of elements and types?
or is that an exercise left for the reader ๐
I'm wrestling with either queueing up blobs of bytes that Python can dissect, or queueing up distinct fields that can be directly accessed
analog bridge
I'm wrestling with either queueing up blobs of bytes that Python can dissect, or...
If parsing in c is significantly faster then in python which I believe it is then this can be left for the user to decide from python side.
crimson ferry
the event queue (like keypad) is ideal conceptually, but my events have relatively large amounts of data
unless maybe the queue contains mini-pointers to pointers to chunks of memory with the data
analog bridge
I am thinking of something like these formatting settings wifi.Packet.RAW, wifi.Packet.SSID, wifi.Packet.RSSI.
crimson ferry
haha, implementing anything like that is many levels above my current (not much) skill ๐
analog bridge
I'll take a look at your implementation today and see if I can help out with something. ๐
crimson ferry
thanks, but I haven't pushed the current work. The existing stuff is really raw and naive (nothing passed up, just serial dump)... what I've been doing today is trying to adapt keypad into a separate module, which was going OK until I got to the ringbuf
analog bridge
That's okay... I will be looking mostly at the forward api.
crimson ferry
i'll push what I have, it's a separate branch 'monitor_queue`
I may also look at how BLE or wifi do scanning, but I've come to think that will give results that are too "chunky"
(bursts of frames followed by intervals of missed frames)
analog bridge
I may also look at how BLE or wifi do scanning, but I've come to think that will...
ya packets are much heavier than that
crimson ferry
(if you look, I just put a couple of fields in to see if it could work and how that would ripple through the code, but that would get unwieldy with a bunch of fields)
I did shared-bindings first, should have started in common-hal
misty storm
Was following the learning guide https://learn.adafruit.com/building-circuitpython/build-circuitpython
But encounter the following error:
$ make V=2 BOARD=circuitplayground_express
GEN build-circuitplayground_express/genhdr/mpversion.h
python3 ../../py/makeversionhdr.py build-circuitplayground_express/genhdr/mpversion.h
make: *** No rule to make target 'freetouch/adafruit_ptc.c', needed by 'build-circuitplayground_express/genhdr/qstr.split'. Stop.
Any advice ?
stuck elbow
@misty storm forgot to do the git submodule update --init?
misty storm
Checked and confirm I have done that. I was able to make mpy-cross.
misty storm
invoke make with the -d flag reveals more hints as below:
Trying implicit prerequisite 'freetouch/SCCS/s.adafruit_ptc.w'.
Trying pattern rule with stem 'adafruit_ptc'.
Rejecting impossible implicit prerequisite 'freetouch/adafruit_ptc.w'.
No implicit rule found for 'freetouch/adafruit_ptc.c'.
Finished prerequisites of target file 'freetouch/adafruit_ptc.c'.
Must remake target 'freetouch/adafruit_ptc.c'.
make: *** No rule to make target 'freetouch/adafruit_ptc.c', needed by 'build-circuitplayground_express/genhdr/qstr.split'. Stop.
So it is saying it needs to create this file adafruit_ptc.c in folder freetouch, which consists of only one file .git
misty storm
<@!515023619118858250> forgot to do the git submodule update --init?
Thanks for the hint. I did carry out this step but always encounter error when trying to clone the esp-idf sub folder. I think this may cause the repo clone to be incomplete and hence the build error.
But then my question is how to tackle this GnuTLS recv error (-24) ?
Cloning into '/mnt/c/Users/kevin/Documents/mcb2/circuitpython/ports/espressif/esp-idf'...
error: RPC failed; curl 56 GnuTLS recv error (-24): Decryption has failed.
fatal: the remote end hung up unexpectedly
fatal: early EOF
fatal: index-pack failed
fatal: clone of 'https://github.com/espressif/esp-idf.git' into submodule path '/mnt/c/Users/kevin/Documents/mcb2/circuitpython/ports/espressif/esp-idf' failed
Failed to clone 'ports/espressif/esp-idf' a second time, aborting
tulip sleet
@misty storm I use this alias:
alias gitsubupdate='git submodule sync --quiet --recursive && git submodule update --init'
so you are using WSL? It looks like you are trying to share storage with your Windows filesystem.
You may also have symbolic-link issues if you try to use /mnt/c
@crimson ferry ports/nrf/common-hal/nrf/_bleio/PacketBuffer.* uses a length header when storing variable-length stuff using ringbuf.
you could also consider using Python lists if you are storing python bojects
stuck elbow
@misty storm I think I've seen someone complaining recently about a certificate error with esp-idf somewhere
tulip sleet
what is the overall thing you are trying to implement?
stuck elbow
looks like one of their certs expired
one possible workaround would be to temporarily disable ssl verification with git config --global http.sslVerify false โ just remember to switch it back to true afterwards
crimson ferry
@tulip sleet thanks, Iโll look at that. Iโm trying to get wifi frames up to pythonโฆ variable-length bytes if raw, and/or a collection of various types of data if parsed in common-hal
tulip sleet
suppose you create a bytes with the data and then just push it on a list
crimson ferry
itโs a continuous process, so the queue is essential
tulip sleet
๐ the first one is, but the "collection of various types of data", that's broken up into objects, is it?
crimson ferry
yes, either way, itโs a continuous firehouse of frames
tulip sleet
that is kind of what PacketBuffer is doing
crimson ferry
iโll dig into _bleio, thanks
onyx hinge
@crimson ferry you could also consider leveraging the code for collections.deque. This would let you work with fully-formed CircuitPython objects rather than low-level C structures. There's not actually a C API for deque objects but we could either (A) expose one internally or (B) your code can go via mp_call_function family functions to call the methods
(I didn't read the full discussion so please forgive if this is irrelevant advice given the full context)
crimson ferry
Thanks, that may well be helpful though most of the stuff between python and simple things in common-hal is still pretty murky to me, so I don't know what a good design looks like. keypad looked really nice to handle continuous events, maybe _bleio.PacketBuffer is a way to extend that to larger / more complex data. There's already a queue at the common-hal level, triggered by an event-handler, but the goal is to get those queued elements containing many bytes (and / or mixed-type data parsed from those bytes) up to Python in a dynamic, more continuous fashion (not like ble or wifi scanning, with a fixed iterable).
tulip sleet
I was thinking of using collections.Deque for keypad events, but it's not included in any builds right now, and I needed to make the code as small as possible, so I went with ringbuf (which is not the greatest piece of code).
I might have also used plain lists, but I needed to make some list operations be public that were static, and I would have had to convert back and forth between Python integers and C integers.
onyx hinge
https://github.com/tiangolo/typer/releases/tag/0.4.0 Typer 0.4.0 supports Click 7 and 8 now!
it looks like cascadetoml and circup don't have a fight anymore, when you install them both. yay!
the new typer came out a month ago but it's the first I noticed.
tulip sleet
hooray!
idle wharf
@lone axle Maybe I missed this, but I was curious if you had a topic planned for your Deep Dive guest host appearance today
lone axle
<@!382939733107408897> Maybe I missed this, but I was curious if you had a topic...
Don't think I mentioned it before so you didn't miss. Plan right now is to try to wire up the chip / LED matrix to test out our first type annotation that has been contributed with this most recent wave of issues: https://github.com/adafruit/Adafruit_CircuitPython_MAX7219/pull/35
Depending on how long that takes (never wired one up before) I may get into some RasPi BlinkaDisplayio testing afterward.
idle wharf
@lone axle Awesome! I'm very curious to see the type annotations, I have not used that much\at all.
lone axle
That is an excellent emoji ๐
supple gale
Same. And I will work on my code as I watch/listen.
lone axle
I'm still very new to typing in python as well
idle wharf
My friend worked on Microsoft Mixer and when they shut it down, he setup his own discord with all the art from theirs ๐
(Where the emoji came from)
lime trellis
@tulip sleet I've been unsuccessful troubleshooting my fake-sleep HardFault_Handler crashes via GDB. It has something to do with GCC optimization but I can't find it even stepping through the entire stack.
what can I try/add to HardFault_Handler to figure out what's causing the issue??
tulip sleet
could you set a breakpoint at all on reset_into_safe_mode or on HardFault_Handler
lime trellis
could you set a breakpoint at all on `reset_into_safe_mode` or on `HardFault_Han...
I can't find it in the disassembly and it doesn't occur when built with DEBUG=1.
(PS it always happens on the 4th iteration of a fake-sleep routine... doesn't occur if I supervisor.reset() that many times, though... and isn't related to the duration of fake sleep)
tulip sleet
you should still be able to set a breakpoint even without DEBUG=1, I think. It still knows the name of global routines.
lime trellis
you should still be able to set a breakpoint even without DEBUG=1, I think. It s...
it can't seem to find it. Is there something I can add to reset_into_safe_mode to prevent it from being optimized out? Maybe a volatile ?
tulip sleet
you want to make it not inline-able
put an MP_NOINLINE in front of the definition
that is a MicroPython macro that has the magic words to make it non-inline
lime trellis
put an `MP_NOINLINE` in front of the definition
Gotcha! trying this now
@tulip sleet I'm working with a feather m4 (D51). Should I be setting up this "micro trace buffer"?
tulip sleet
I think that's only SAMD21
lime trellis
@tulip sleet finally got it to break on reset_into_safe_mode. What sort of things can I query to determine my issue?
tulip sleet
jsut do a bt (backtrace) to see what was happening when it hit that
lime trellis
This is all I get. Should I investigate mp_load_method_maybe?
(gdb) bt
#0 0x00044cde in reset_into_safe_mode ()
#1 0x000053c4 in HardFault_Handler ()
#2 <signal handler called>
#3 0x0002c92a in mp_load_method_maybe ()
#4 0x2001e0e0 in ?? ()
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
ugh, could be better
so damaged stack already. you could populate the code around where you think things are going bad with mp_printf(&mp_plat_print, ..)s and do some divide and conquering to narrow where it might be going wrong
setting a breakpoint on mp_load_method_maybe() is not going to help because it's called too often
lime trellis
so damaged stack already. you could populate the code around where you think thi...
I've been using printf to hunt for things but at this point I'm all out of stuff I can try.
I do know it's always the 4th fake-sleep cycle during a call to alarm.exit_and_deep_sleep_until_alarms(time_alarm). My Gut is that it's something to do with this portion of main.c
#if CIRCUITPY_ALARM
if (result.return_code & PYEXEC_DEEP_SLEEP) {
// Make sure we have been awake long enough for USB to connect (enumeration delay).
int64_t connecting_delay_ticks = CIRCUITPY_USB_CONNECTED_SLEEP_DELAY * 1024 - port_get_raw_ticks(NULL);
// Until it's safe to decide whether we're real/fake sleeping
if (fake_sleeping) {
// This waits until a pretend deep sleep alarm occurs. They are set
// during common_hal_alarm_set_deep_sleep_alarms. On some platforms
// it may also return due to another interrupt, that's why we check
// for deep sleep alarms above. If it wasn't a deep sleep alarm,
// then we'll idle here again.
common_hal_alarm_pretending_deep_sleep();
} else if
so if you just print something before every statement there and see when one doesn't show up?
lime trellis
so if you just print something before every statement there and see when one doe...
I've drilled down to exactly when it happens. But I'll try and dig deeper.
lime trellis
so if you just print something before every statement there and see when one doe...
managed to get a little bit farther? Should I look into a gc issue?
#0 0x00044d06 in reset_into_safe_mode ()
#1 0x000053c4 in HardFault_Handler ()
#2 <signal handler called>
#3 0x2001d720 in ?? ()
#4 0x00022830 in gc_collect_end ()
#5 0x200015e0 in _mscd_itf ()
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
it's probably worth checking the code to make sure that any objects you have stashed away are being marked in gc. If you made a static that contains an object, it will need to be made part of gc.
atomic summit
Hey, was @lone axle streaming today?
lone axle
Hey, was <@!382939733107408897> streaming today?
starting right now
tulip sleet
he is going to, see discussion in #live-broadcast-chat
lime trellis
it's probably worth checking the code to make sure that any objects you have sta...
I didn't create any new static functions (beyond what TimeAlarm already has in most ports), but I did REMOVE a static ( timer_callback ) and just made it a void timer_callback(void).
Where is this list of special gc objects? I'm combing through /py/gc.h now
tulip sleet
I didn't create any new static functions (beyond what TimeAlarm already has in m...
Not functions, but variables. We are talking about the root pointers and the variables gcโd in the various *_gc_collect routines
lime trellis
Not functions, but variables. We are talking about the root pointers and the var...
I see! Ok I'll investigate and learn more. thank you!!
manic glacierBOT
misty storm
<@515023619118858250> I use this alias:
`alias gitsubupdate='git submodule sync ...
Thank you. I moved to build on another Ubuntu (not Ubuntu on WSL) and the build completes successfully!
I try to test the firmware but found that the "version" as printed on the REPL is "2025d0c", which happens to be the beginning digits of the commit as reported on the circuitpython github (the commit number ??)
https://github.com/adafruit/circuitpython/commit/2025d0c0605c4c0b21f58e1337e35546e011668b
Now when I try to update the libraries with circup, it seems confused by this "version" number and comes back with a "there was a problem". How can one controls the version as I should have put it as 7.0 ?
misty storm
skip the cicrup and move the libs manually. I still bump into this
Adafruit CircuitPython 2025d0c on 2021-10-02; MakerDiary nRF52840 MDK USB Dongle with nRF52840
from adafruit_ble import BLERadio
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "adafruit_ble/init.py", line 17, in <module>
ImportError: This release is not compatible with CircuitPython 4.x; use library release 1.x.xAny advice ?
tulip sleet
Thank you. I moved to build on another Ubuntu (not Ubuntu on WSL) and the build ...
Have you update circup recently? I didn't see this problem when I used circup recently on a build with such a version
tulip sleet
skip the cicrup and move the libs manually. I still bump into this
Adafruit Cir...
Aha, it looks like you are building off master instead of main?
misty storm
Aha, it looks like you are building off master instead of main?
Yes. May I ask how to build from main ? I just follow your learning guide literally.๐
tulip sleet
if you type git status, what does it say?
misty storm
skip the cicrup and move the libs manually. I still bump into this
Adafruit Cir...
But if I put circuitpython 7.0 back, then things work as before as it should.