#circuitpython-dev
manic glacierBOT
sharp jacinth
2 mic inputs and one AUX input
tulip sleet
which circuitpython boards do you have?
do have any SAMD21 or SAMD51? They have DAC's, and you can test them into the amp.
sharp jacinth
I may at home, I'd need a headphone jack right?
tulip sleet
@onyx hinge I am wondering if the CS timing is marginal, etc.
tulip sleet
I may at home, I'd need a headphone jack right?
you just need a cable that plugs into the amp and some alligator clip leads. Is it 1/4" in or 1/8" in or RCA phono jack?
onyx hinge
@tulip sleet yeah there's no particular delay from CS going low to the start of the SPI traffic
in the python code I transcribed that wouldn't matter
tulip sleet
you mean any delay is really short? Maybe adding a little delay of pulling CS low would make a diff?
i will test that later (or look at the waveforms)
sharp jacinth
you just need a cable that plugs into the amp and some alligator clip leads. Is ...
1/8th I think. 1/4th is the size that goes into instruments I think
sharp jacinth
I did haha until I cut it, I can get another.
onyx hinge
@tulip sleet your other error is "couldn't determine SD card version"?
tulip sleet
well, you could strip the wires on the cut end, but if it's not urgent I would say get another. You need mini alligator clip leads too, or you need a jack
tulip sleet
<@!329766224093249548> your other error is "couldn't determine SD card version"?
yes, and you see that the data printed is all zeros, which is really odd
i haven't gotten "no SD card" on the breakout at all
and weirdly, as i mentioned, one that failed on the GC4, works on the MM4 with breakout, but another that works on the GC4 gets the version error on the MM4
no wait, I think that's wrong
I have little boxes with SD cards in them and notes. They all look alike 🙂
onyx hinge
I think it's less likely that it's a CS timing issue, I think that CS is asserted (false level) all the way from the start of detect to where CMD 8 is ussued
weirdly this document never talks about the time from CS to data, except that you have to clock out N_CS "H" bits. N_CS is min 0, max —, in units of 8 clock cycles (even though 4 cycles are shown above)
onyx hinge
that's what "wait for idle" does, it it clocks out 0xffs until it gets an 0xff back in
these cards are black magic to me 😕
tulip sleet
I didn't know you had to toggle CS like that. I might try the breakout with shorter wires
or maybe an adalogger featherwing
onyx hinge
another weird CI fail on my PR .. https://github.com/adafruit/circuitpython/pull/5688 "fatal: No tags can describe '8d2a4168454ae5c7786a478fd8876c8bbfe0ec37'." it's targeted to 7.1.x, maybe that matters? @slender iron did you arrive at an understanding of this?
or is that what your just merged PR is about?
slender iron
ya, I fixed espressif in a follow up
onyx hinge
thank you!
slender iron
I could have sworn I had grepped for each spot
onyx hinge
is the best thing to do create an empty commit?
onyx hinge
interesting, I didn't know and didn't want to guess
slender iron
you can check to see if that command has a new sha / is correct in a rerun
onyx hinge
@slender iron you'd rather just not merge https://github.com/adafruit/circuitpython/pull/5678 in any form?
onyx hinge
OK. does it matter for your purposes that espressif didn't want to allocate different PIDs for the two versions of the esp32-c3 board (psram / no psram)?
we can just allow that duplicate PID right?
slender iron
ya, we can allow it
onyx hinge
okay, great
crimson ferry
lol, that's concurrent execution for you
onyx hinge
oof sorry if I'm stepping on your toes!
onyx hinge
Not exactly circuitpython-dev material, until it is. In a personal python project, I had pre-commit suddenly start failing even though nothing had changed. turns out what changed was the github default Python3 version (installed by actions/setup-python), as far as I can tell. Whatever exactly happened, the consequence was a new pylint diagnostic, wwvb/updateiers.py:90:19: E0602: Undefined variable 'wwvb_start' (undefined-variable) -- now I specify a version of python to install within github actions for pre-commit, but of course developers might have different python versions themselves ... 😕
(not that anyone develops this project with me, mind you)
teal thorn
I'm looking to add some new NXP ports and looking for some advice. Can we rename the mimxrt10xx directory to NXP since these share a common SDK, or is it better to add a parallel nxp directory? I may start in a new nxp directory using the official SDK git repo, but I will need to copy most of the files from the mimxrt10xx port to start. Does anyone have any suggestions before I start?
sharp jacinth
@tulip sleet I should clarify, I'm blaming the stuttering in audio I experience with the pi on the OS/Pygame. I'm not using baremetal cp as I deployed this product before that was available. I think the audio is just cutting out now and again because the OS is busy. I look forward to trying baremetal cp on the pi soon though
Just wanted to make sure I didn't miscommunicate
manic glacierBOT
jaunty juniper
it's an LED that is blue
the xiao M0 doesn't have an RGB LED
but I mean, there isn't a strong rule, it's just RED_LED looks to me like it designates a red LED, while LED_R designates the R of the RGB LED
viscid pine
it could be argued that it's best to keep it more compatible with the xiao m0, since it's already using the xiao pin naming (D0, D1 ...)
manic glacierBOT
viscid pine
now i'm wondering why the m0 doesn't have pin defs for the tx and rx leds
i can't decide @jaunty juniper
jaunty juniper
so, I don't have the board, what you are saying according to your PR is that there isn't a neopixel on the board, but an RGB LED with 3 PWM component ?
jaunty juniper
ahhh ok
viscid pine
the old code was referring to stuff in seeed's schematic, which isn't right
I guess they were originally going to make it with the same LEDs as the xiao m0, then decided to do RGB instead
jaunty juniper
yeah the way you named the pins makes sense to me now
viscid pine
i should contact them about that schematic
jaunty juniper
hey, when cloning Circuitpython main, initing the submodules with: git submodule sync --quiet --recursive && git submodule update --init pulls a whole whooping 23 GB for ports/broadcom/firmware would there be a trick to avoid that ?
analog bridge
hey, when cloning Circuitpython main, initing the submodules with: `git submodul...
make fetch-submodules was recently added for this
https://github.com/adafruit/circuitpython/issues/5619
jaunty juniper
`make fetch-submodules` was recently added for this
https://github.com/adafruit/...
ah nice, thank you 👍
solar whale
@idle owl FYI -- THe PyLeap App has been updated, and I tried some of the new demos but I can't get them to load - I tried the Light Meter and Sound Meter. They start to load then hang - the CPB starts to blink red every 5 or so secs. I looked a the files with File Glider and I can see that the new code.py has been loaded, but not the rest of the files. the older demos like Blink and Glide Rainbow still load and run OK. I opened an issue on the PLeap-IOS repo, but I thought you's be interested. Have you tried it recently. I am using the latest build of CPB.
tulip sleet
@onyx hinge Have you noticed that the linked issues stuff does not seem to be linking or closing issues? I have seen this in a number of recent PR's. Maybe it is because the PR's are not on main.
onyx hinge
that might be why
manic glacierBOT
onyx hinge
"Successfully merging this pull request may close these issues." that's confidence-inspiring 🙂
solar whale
Will it also get in to main?
onyx hinge
Note: The special keywords in a pull request description are interpreted when the pull request targets the repository's default branch. However, if the PR's base is any other branch, then these keywords are ignored, no links are created and merging the PR has no effect on the issues. If you want to link a pull request to an issue using a keyword, the PR must be on the default branch.
https://docs.github.com/en/issues/tracking-your-work-with-issues/linking-a-pull-request-to-an-issue
tulip sleet
aha, thanks. I guess the idea is that the issue isn't necessarily fixed until the fix is in the main branch. I can understand being annoyed if it were just fixed in a dev branch or even a PR to a fix branch
onyx hinge
that doesn't quite describe what happened, because the link was created, but merging the PR didn't affect the state of the issue
tulip sleet
I have seen the link not created also.
onyx hinge
ahh in this case I created the PR to main and then changed what they call its "base"
using the webui
tulip sleet
Will it also get in to main?
Scott just PR'd a bunch of 7.1.x changes to main, and we will do that again on a frequent basis. This seems like an important one.
solar whale
Yeas -- I saw that -- too bad this one missed the cut....
tulip sleet
is it easy for you to test the SD/ESP32SPI issue on 7.1.x? (Do you have time?) Or I can make another merge PR.
solar whale
I can test it on 7.1x, but my previous testing was with 7.2 so I wanted to keep it the same.
I should be able to pull it in myself to try it.
tulip sleet
@onyx hinge so you can repoint an existing PR to merge against a different branch? I would want to do that for https://github.com/adafruit/circuitpython/pull/5693, for example. I don't see that I can do it, but I may be missing where. Or maybe the submitter needs to do it.
solar whale
hmm -- the pr has different version of several submoduled from 7.2 -- not sure if I should fetch tem or not ```remote: Enumerating objects: 10, done.
remote: Counting objects: 100% (10/10), done.
remote: Compressing objects: 100% (4/4), done.
remote: Total 10 (delta 6), reused 10 (delta 6), pack-reused 0
Unpacking objects: 100% (10/10), 2.73 KiB | 465.00 KiB/s, done.
From https://github.com/adafruit/circuitpython
- [new ref] refs/pull/5688/head -> pr_5688
jerryneedell@jerryneedell-ubuntu-macmini:~/projects/circuitpython$ git checkout pr_5688
M data/nvm.toml
M ports/broadcom/firmware
M ports/broadcom/peripherals
M ports/espressif/esp-idf
Switched to branch 'pr_5688'
@tulip sleet there's a thing you can click, but it's only part of the story, because the ref where the PR starts needs to be appropriate... anyway, on an open PR, click the "edit" button that is used for editing the PR title and then the base becomes editable
so to do it on my end I also rebased & force-pushed my changes, which would be up to the PR originator..
tulip sleet
aha, great, I will try that. That PR is very simple. I can flip it back if it looks bad.
oh, ugh, it was bad, I see; I changed it back
ok, never mind.
onyx hinge
yeah...
tulip sleet
ty ^^
onyx hinge
I guess it would automatically pick up any other 7.1.x additions until it is merged, just like any other PR...
main furnace
I'd like to try the sdcard fix on the feather m4 express. Do I need the 7.2 build or 7.1? My sd card is on the TFTFeatherWing 3.5.
solar whale
Is this expected -- I was going to try pr 5694 --- why the new submodules ```jerryneedell@jerryneedell-ubuntu-macmini:~/projects/circuitpython$ git fetch origin pull/5694/head:pr_5694
remote: Enumerating objects: 11, done.
remote: Counting objects: 100% (11/11), done.
remote: Compressing objects: 100% (5/5), done.
remote: Total 11 (delta 6), reused 10 (delta 6), pack-reused 0
Unpacking objects: 100% (11/11), 3.35 KiB | 3.35 MiB/s, done.
From https://github.com/adafruit/circuitpython
- [new ref] refs/pull/5694/head -> pr_5694
jerryneedell@jerryneedell-ubuntu-macmini:~/projects/circuitpython$ git checkout pr_5694
M data/nvm.toml
M ports/broadcom/firmware
M ports/broadcom/peripherals
M ports/espressif/esp-idf
Switched to branch 'pr_5694'
the submodules are not in sync between 7.1.x and main right now. You can just git submodule update --init
It will not fetch the whole broadcom/firmware thing
maybe git submodule sync also
solar whale
git submodule update --init
oops wrong screen
willt that use the correct submodules for 7.2 and not 7.1.x?
tulip sleet
it will use the right submodules for whatever the git tree says?
solar whale
I am trying 5694 on 7.2
tulip sleet
i think it will be ok if you are not going to be testing on espressif or broadcom it doesn't really matter
solar whale
true -- it still ends up with ```erryneedell@jerryneedell-ubuntu-macmini:~/projects/circuitpython$ git diff
diff --git a/ports/espressif/esp-idf b/ports/espressif/esp-idf
--- a/ports/espressif/esp-idf
+++ b/ports/espressif/esp-idf
@@ -1 +1 @@
-Subproject commit e493a4c30e1b2f7c54a376327388b9e12262cbbf
+Subproject commit e493a4c30e1b2f7c54a376327388b9e12262cbbf-dirty
you can update so you don't see that and then go back again
solar whale
pr 5694 is working fine with sdcardio on my pigamer with the ariflift feather wing and a variety of different SDCards
tulip sleet
was it not working before with the same combos?
I can't remember whether you could reproduce that or not
solar whale
At mylat test, it was also working -- I did not know why it started working.. I never had any issue with the esp32spi -- only the SDCards
I am "hot swaping" the SDCards and doing a soft reboot -- still working
manic glacierBOT
tulip sleet
@onyx hinge did you cancel the previous merge build just to save time? Fine with me.
solar whale
previously I had SDCard issues with 7.1x -- not 7.2
onyx hinge
@tulip sleet no...?
tulip sleet
hunh a run got cancelled. https://github.com/adafruit/circuitpython/actions see #7517
"request to deprovision" I have never seen that before. Maybe a CI glitch
analog bridge
hunh a run got cancelled. <https://github.com/adafruit/circuitpython/actions> se...
I might have seen that before, I think its a github bug.
tulip sleet
well, it will save time, since it's superseded anyway
solar whale
If I check out 7.1.x -- it does not show the latest merge -- I don't think.
tulip sleet
@analog bridge if you are on, did you see https://github.com/adafruit/circuitpython/issues/5682
solar whale
M data/nvm.toml
M ports/broadcom/firmware
M ports/broadcom/peripherals
M ports/espressif/esp-idf
Note: switching to 'origin/7.1.x'.
You are in 'detached HEAD' state. You can look around, make experimental
changes and commit them, and you can discard any commits you make in this
state without impacting any branches by switching back to a branch.
If you want to create a new branch to retain commits you create, you may
do so (now or later) by using -c with the switch command. Example:
git switch -c <new-branch-name>
Or undo this operation with:
git switch -
Turn off this advice by setting config variable advice.detachedHead to false
HEAD is now at cd47941c7 Merge pull request #5687 from tannewt/more_shas
is origin you or adafruit?
solar whale
adafruit -- I am not on my fork
tulip sleet
i am not sure why you would see that, it's a branch, not a tag
if you just git checkout 7.1.x does it make any difference?
or you need to pull
solar whale
maybe ```origin https://github.com/adafruit/circuitpython.git (push)
jerryneedell@jerryneedell-ubuntu-macmini:~/projects/circuitpython$ git checkout 7.1.x
M data/nvm.toml
M ports/broadcom/firmware
M ports/broadcom/peripherals
M ports/espressif/esp-idf
Branch '7.1.x' set up to track remote branch '7.1.x' from 'origin'.
Switched to a new branch '7.1.x'
ah -- then a git pull brings it in
remote: Enumerating objects: 24, done.
remote: Counting objects: 100% (24/24), done.
remote: Compressing objects: 100% (11/11), done.
remote: Total 24 (delta 13), reused 18 (delta 13), pack-reused 0
Unpacking objects: 100% (24/24), 10.50 KiB | 1.17 MiB/s, done.
From https://github.com/adafruit/circuitpython
cd47941c7..8bd35d954 7.1.x -> origin/7.1.x
8eee16c7a..3bdf2a5cd main -> origin/main
Updating cd47941c7..8bd35d954
Fast-forward
shared-module/sdcardio/SDCard.c | 10 +++++++++-
1 file changed, 9 insertions(+), 1 deletion(-)
ok -- now I'll build and try it on the pygamer
solar whale
unfortunately, the released 7.1.0.-beta.1 is also now working for me....
tulip sleet
ok, we need to ask the other people with this issue if the latest ones work for them now
solar whale
FWIW -- 7.1.x works as well with the PR
analog bridge
<@!747673663419187281> if you are on, did you see <https://github.com/adafruit/c...
Can't draw any conclusions from the quick look I just took at it, will look at it in detail sometime later, it seems that the author is unable to replicate it now.
tulip sleet
thanks for looking
main furnace
Tried Adafruit CircuitPython 7.2.0-alpha.0-85-g3bdf2a5cd on 2021-12-09; Adafruit Feather M4 Express with samd51j19.
Result: No SD card found.
-- It does work with the 7.0.0 release.
tulip sleet
Tried Adafruit CircuitPython 7.2.0-alpha.0-85-g3bdf2a5cd on 2021-12-09; Adafruit...
The fix hasn't been merged into main yet. It is on the 7.1.x branch. Try the artifacts available here to test: https://github.com/adafruit/circuitpython/actions/runs/1557003762
lone sandal
main furnace
Okay it is working on Adafruit CircuitPython 7.1.0-beta.1-61-gcc4e7f868 on 2021-12-09. Now I can save more screenshots!
manic glacierBOT
orchid basinBOT
slender iron
I'm looking to add some new NXP ports and looking for some advice. Can we renam...
If it's all the same SDK then I think renaming is fine. What chip series are you adding?
orchid basinBOT
idle owl
<@!330227457296957440> FYI -- THe PyLeap App has been updated, and I tried some ...
I'm getting the same result.
solar whale
WEll, I guess that's good....
idle owl
I'll ping Trevor. The "Downloaded files" looks suspect, but I don't know how it lists things. It has all the files listed separately from the directory, and maybe that's by design, or maybe the Bundle isn't right.
idle owl
I created the issue and left feedback on the guide Trevor just posted.
It works on 7.0.0.
So something is up with latest. beta1 included
solar whale
So something is up with latest. beta1 included
ruh roh 😉
idle owl
Heh yeah. I told Trev to ping Scott because I'm of no use here. I replied to the issue and said I replicated it on the beta but not on stable.
solar whale
Thanks
idle owl
You're welcome! Thanks for reporting it.
slender iron
I just pinged trev about it on iosdev too
idle owl
I just pinged trev about it on iosdev too
Sounds good.
slender iron
I read the channels in the wrong order 🙂
idle owl
There's a GItHub issue for it as well.
slender iron
ya, I saw that but no one had responded when I looked
idle owl
Fair enough. Updated now!
I asked Trev to reply to it as well, but I know he's not as comfortable with GitHub.
slender iron
thanks for helping him with it
solar whale
I just confirmed the same behavior with the latest tip of main 7.2...
slender iron
I asked trev to make a cp issue and I'll look to fix it for 7.1
anyone with a rp2040 itsy want to confirm https://github.com/adafruit/circuitpython/issues/5685?
solar whale
7.0.0 fails for me as well -- with Sound Meter anyway
idle owl
7.0.0 fails for me as well -- with Sound Meter anyway
I didn't try that one... hmm.
manic glacierBOT
idle owl
7.0.0 fails for me as well -- with Sound Meter anyway
Worked for me.
solar whale
so does Light Meter, hmmmm
slender iron
is the drive already full?
solar whale
Il'' try eraseing
slender iron
maybe its not handling the case of running out of space
solar whale
nevermind -- I hade it l=plugged into my computer...USB
slender iron
it does but it won't remove unused files
idle owl
nevermind -- I hade it l=plugged into my computer...USB
😂 I did the same thing when testing for Trevor 20 minutes ago.
slender iron
does it not have an error when on usb?
idle owl
does it not have an error when on usb?
No.
solar whale
it just hangs
idle owl
^^
slender iron
I thought we had added that
maybe just in file glider
CP definitely returns an error code to the client
specific to usb being connected
solar whale
still not working for me 😦
idle owl
anyone with a rp2040 itsy want to confirm <https://github.com/adafruit/circuitpy...
If no one else can, I can, but I'm not sure what I'm looking for. Seems a little out of my wheelhouse. If it requires a Saleae, I can do it with assistance, but I don't really have a scope.
solar whale
no - I did storage.erase_filesystem()
slender iron
no worries @idle owl
slender iron
I think I had to send the glasses one twice when I did it
solar whale
glasses transfers OK but does not work on the CPB
solar whale
it does the 2 red blinks like it did when the other trasnfers failed
slender iron
2 red blinks is a error probably due to a partial transfer
you can do storage.disable_usb_drive() in boot.py to get usb serial output plus BLE
solar whale
Ok -- will try that in a bit.
slender iron
thank you
solar whale
after disabling the USB drive -- I loaded blink. it worked ok -- tried loading sound meter and I get ```Adafruit CircuitPython 7.0.0 on 2021-09-20; Adafruit Circuit Playground Bluefruit with nRF52840
soft reboot
Auto-reload is on. Simply save files over USB to run them or enter REPL to disable.
code.py output:
Traceback (most recent call last):
File "code.py", line 10, in <module>
ImportError: no module named 'adafruit_circuitplayground'
Code done running.
Press any key to enter the REPL. Use CTRL-D to reload.
which agrees with a partial load
@idle owl must have a magic touch....
BTW -- I have IOS 15.1 on my iPhone
slender iron
@solar whale did the sound meter actually finish the "please wait" prompt?
solar whale
No -- just hangs partway through
slender iron
ok, I see that too on both 7.0 and main
solar whale
slender iron
yup, I see that too
solar whale
Glad I'm not the only one!
idle owl
<@!330227457296957440> must have a magic touch....
😄 Maybe?
slender iron
@idle owl did you already have some folders on your drive?
I'm not sure the ios app is making folders before trying to write to them
idle owl
<@!330227457296957440> did you already have some folders on your drive?
Blink works, but the others do not now.
teal thorn
If it's all the same SDK then I think renaming is fine. What chip series are you...
Considering K32L2B at first, but LPC55 and RT1170 may follow
slender iron
makes sense to me. anything that needs a folder won't work
slender iron
Considering K32L2B at first, but LPC55 and RT1170 may follow
That is much tinier than the imx rt. it might be worth doing separately. it's more like a samd21
probably not worth doing due to limited ram. we are trying to dissuade people away from samd21
lone axle
Is there a CircuitPython driver that works with: https://www.adafruit.com/product/5146 ? I don't see it mentioned on the product page so I am guessing not. But I did notice the default I2C address is 0x50 and found that is the address used by default in the Adafruit_CircuitPython_FRAM
library for I2C connected devices. I'm not really familiar with FRAM and EEPROM are those different enough that they would need different drivers? I see the arduino driver linked covers both but has different classes for them.
teal thorn
That is much tinier than the imx rt. it might be worth doing separately. it's mo...
They use the same SDK and many of the peripheral libraries/hal will be compatible, but some files may end up more family specific. Is there a good way for families to share a common SDK git repo?
slender iron
@teal thorn I don't think it's super clear either way. Sharing the same port directory should be fine
within a port we use ifdefs to switch chip family
teal thorn
I'll probably create a new nxp directory for this work. Then we can migrate the mimxrt parts to the updated SDK later.
slender iron
sure, that works too
viscid pine
nothing like a morning fight with git
stuck elbow
who won?
viscid pine
it's ongoing
stuck elbow
it's like sparring with a pig in mud: halfway through you realize the pig enjoys it
idle owl
@tulip sleet TimeoutError: Clock stretch too long on RP2040 with NeoSlider. You had to make a build for JP, right? Is it in latest on S3, or still in the magical world of Dan only?
Or is this something different
tulip sleet
<@!329766224093249548> `TimeoutError: Clock stretch too long` on RP2040 with Neo...
It has been merged, so an Absolute Newest or the latest 7.1.x build will work.
idle owl
It has been merged, so an Absolute Newest or the latest 7.1.x build will work.
Thank you.
tulip sleet
that is the error you would get
viscid pine
@tulip sleet you requested that I PR against 7.1.x, but the xiao rp2040 isn't in the branch yet. How should I proceed?
idle owl
that is the error you would get
Thank you, latest worked. Though I had to unplug it and plug it back in, got a no pullups error on the first go. Shrug. Works now.
tulip sleet
<@!329766224093249548> you requested that I PR against 7.1.x, but the xiao rp204...
ah, I thought it was, and I even checked, but it's in main. So never mind 🙂 main is good.
tulip sleet
@slender iron @idle owl I think I will make a 7.1.0-beta.2 tomorrow, since we have several fixes that people are expecting (NeoSlider, SD card fix, bleio HCI, board fixes, etc.). Draft release notes are done. I also started draft notes (not yet on GitHub) for the first 7.2.0-alpha (or maybe it will be beta) release, because it's easier to add PR's to those release notes now as I review the PR's in the last few weeks.
slender iron
Sounds good. Thank you!
manic glacierBOT
idle owl
<@!252717193496756235> <@!330227457296957440> I think I will make a 7.1.0-beta.2...
Sounds great! Thanks!
solar whale
<@!252717193496756235> <@!330227457296957440> I think I will make a 7.1.0-beta.2...
Is it a simple matter to update the "frozen"modules before doing the release? rfm69 and rfm9x have been updated. Not sure about others.
tulip sleet
Is it a simple matter to update the "frozen"modules before doing the release? rf...
Yes, very simple, I can do that right now.
thanks for the reminder. It should just be part of the release process.
timber mango
hi, I tried to build circuitpython on ubuntu to add board Elecfreaks pico:ed (microbit edge) but of course, I am not maker of this board and was stopped by usb pid thing ... all the work is documented here, but may be come clarification note for such efforts in the begining of "adding board doc" will be better? how its OK to try such things? it seems that manufacturer of this board is not responing and I am not sure of usb vid/pid for this - as I initially flashed generic rpi pico firmware there, it reported theirs usb vid/pid on COM port in windows - may be that with stock micropython fw it will report something other? I doubt, I remember it also had rpi pico micropython / uf2 bootloader probably... my work is here: https://github.com/elecfreaks/pico_ed/issues/1
when I left entire usb config as-is for generic rpi pico, cp was built without any duplicity report, so here I dont understand properly this
tulip sleet
hi, I tried to build circuitpython on ubuntu to add board Elecfreaks pico:ed (mi...
is your issue mostly deciding pin mapping?
timber mango
is your issue mostly deciding pin mapping?
umm, dealing with usb PID mostly - I already did P0-P16, P19/SCL P20/SDA (comparing pinouts of microbit, CLUE ... here is not well placed compatible hw SPI only) ... its working but with raspberry pi pico usb settings .... everything is in attached zip, together with IS31FL3731 new pico_ed module
tulip sleet
umm, dealing with usb PID mostly - I already did P0-P16, P19/SCL P20/SDA (compar...
if the mfr does not have their own VID (costs a few $k), Raspberry Pi gives out PIDs: https://github.com/raspberrypi/usb-pid
timber mango
if the mfr does not have their own VID (costs a few $k), Raspberry Pi gives out ...
ya, studied entire doc, I notified them about that, but they are silent...
tulip sleet
ya, studied entire doc, I notified them about that, but they are silent...
I looked at their site: they sell add-on's, but the base board is a micro:bit, right?
I cannot find the pico:ed product
timber mango
I looked at their site: they sell add-on's, but the base board is a micro:bit, r...
no no, its this rp2040 based board https://www.elecfreaks.com/picoed.html
timber mango
I cannot find the pico:ed product
but even theirs pinout has mistake in buttons, everything is in my first github attached zip package
tulip sleet
how are you trying to communicate with them?
could try twitter, etc.
but if they are uninterested, it is their loss
timber mango
could try twitter, etc.
will try twitter too, yes, okay
slender iron
@timber mango I've given out PIDs under pid.codes for open source hardware on closed source hardware
tulip sleet
afk - back in a while and will do the 7.1.0-beta.2 blog and forum posts
slender iron
thanks dan!
tulip sleet
the circuitpython.org PR action failed yet again. adabot Python code not available, though it is cloned (but not installed via requirements??)
i did it by hand yet again. This time I wrote down how I did it
slender iron
I can take a peek
lime trellis
so far sd card fix is working great on 7.2. Thx @tulip sleet !
tulip sleet
so far sd card fix is working great on 7.2. Thx <@!329766224093249548> !
Thanks @onyx hinge too. I provided the testing and bisect, and Jeff fixed it.
proven garnet
How are the hex numbers for ManufacturerDataFields determined in this library? I want to start getting into BLE libraries, and there's an issue to add sound to this library but I'm not sure where to start, if it's even possible at this point: https://github.com/adafruit/Adafruit_CircuitPython_BLE_BroadcastNet/blob/main/adafruit_ble_broadcastnet.py
Asking specifically for the AdafruitSensorMeasurement class
tulip sleet
How are the hex numbers for ManufacturerDataFields determined in this library? ...
The Adafruit company ID we paid for. The other numbers are "made up"; they define a standard for this kind of advertisement.
proven garnet
Thanks for the explanation! Is there a place where info on what currently makes up this working standard is stored?
tulip sleet
https://circuitpython.readthedocs.io/projects/ble_broadcastnet/en/latest/api.html, but the numbers are not there
solar whale
Are there any web server demos for the esp32s2? came up in a forum question if anyone wants to respond. https://forums.adafruit.com/viewtopic.php?f=53&t=186099
proven garnet
https://circuitpython.readthedocs.io/projects/ble_broadcastnet/en/latest/api.htm...
So how would one go about adding one, like for sound, as one of the issues mentions? Just pick an arbitrary int?
tulip sleet
So how would one go about adding one, like for sound, as one of the issues menti...
Yes, and if you feel like it, submit a PR
they are approximately consecutive, I think
Or if you want yours not to interfere, pick something way out of range
proven garnet
Yes, and if you feel like it, submit a PR
Sweet! Thanks, I'll get on that in the next few days. Why are some listed as not supported then (like datetimes) though? Is something preventing implementations of those?
tulip sleet
there may just not be any library code written for some of them. I'm not sure. Scott is the one who worked on this
proven garnet
Thank! Appreciate the explanation!
crimson ferry
Are there any web server demos for the esp32s2? came up in a forum question if a...
⤴️ @jaunty juniper I believe has something https://github.com/Neradoc/circuitpython-native-wsgiserver
manic glacierBOT
idle owl
@lone axle I trust your judgement. If you think a PR should be closed, you can close it. Worst case, it can always be reopened with the click of a button.
You can still tag me if you like so it's on the record and more readily available in my email.
lone axle
<@!382939733107408897> I trust your judgement. If you think a PR should be close...
okay sounds good. Thank you. Looking through a few more tonight as well. I'll close them up if no movement for a long while.
blissful pollen
@idle owl Here is the branch for what I was thinking of for the native IS31FL3741 support helper libraries. No rush I know it is the weekend.
idle owl
okay sounds good. Thank you. Looking through a few more tonight as well. I'll cl...
There are a number like that. You can also tag the original author one more time if you feel it makes sense, make a note for yourself, and give them a couple of weeks.
The oldest PR we have was still being worked on every so often by the author, for example. Worth tagging them when you get to it. It's in a repo that starts with T, so end of the list somewhere.
manic glacierBOT
idle owl
<@!330227457296957440> Here is the branch for what I was thinking of for the nat...
Great! Thank you. I've added it to my list. I'll get back to you when I get a chance to take a look.
meager adder
I'm trying to build circuitpython for the raspberry pi pico from 7.0.0 in WSL. It's returning a TypeError.
> git checkout 7.0.0
HEAD is now at 42523a183 Merge pull request #5357 from dhalbert/revert-5341
> git status
HEAD detached at 7.0.0
nothing to commit, working tree clean
> cd ports/raspberrypi
> make V=1 BOARD=raspberry_pi_pico
GEN build-raspberry_pi_pico/genhdr/mpversion.h
CC ../../py/builtinhelp.c
CC ../../py/modsys.c
CC ../../main.c
CC supervisor/port.c
GEN flash_info.h.jinja
Traceback (most recent call last):
File "gen_stage2.py", line 96, in <module>
typer.run(main)
File "/home/me/.local/lib/python3.8/site-packages/typer/main.py", line 864, in run
app()
File "/home/me/.local/lib/python3.8/site-packages/typer/main.py", line 214, in __call__
return get_command(self)(*args, **kwargs)
File "/home/me/.local/lib/python3.8/site-packages/typer/main.py", line 239, in get_command
click_command = get_command_from_info(typer_instance.registered_commands[0])
File "/home/me/.local/lib/python3.8/site-packages/typer/main.py", line 427, in get_command_from_info
command = cls(
TypeError: __init__() got an unexpected keyword argument 'no_args_is_help'
make: *** [Makefile:265: build-raspberry_pi_pico/genhdr/flash_info.h] Error 1
Can I get some help troubleshooting this issue?
viscid pine
looks like the typer library might be out of date?
try pip install --user --upgrade typer
meager adder
try `pip install --user --upgrade typer`
That seems to have done the trick. The build is progressing now. Thank you!
lone axle
It seems like there was something in commit 8cf269d2a which has caused some trouble. On a PyPortal I'm noticing that starting with that build displayio examples are no longer working properly and going to REPL and running commands is not printing the >>> and commands that are run onto the screen.
Dexter Starboard tested the same things and found that on PyBadge displayio does not have the same problem. But that starting with that same build the auto-reload functionality stops working, so saving a file change does not re-run code.py.
Possibly these are two different symptoms of the same root cause. There are some more details in the issue linked above this message.
Well the name of the build has 8cf269d2a in it. I'm not sure if that correlates exactly with the commit hash.
tough flax
@slender iron - watching your deep dive... on line 56 of your neopixel code, do you have an extra "=" in that operator? My eyes are old, but it looks like you have a "<<==" instead of "<<=". Not sure if that's intentional (or why that builds). Just thought I'd share that it caught my eye.
manic glacierBOT
manic glacierBOT
lone axle
I wonder if it would be helpful to make some code that parses the pins.c file in each device in the circuitpython repo and generates a simulated screenshot of what running
import board
dir(board)
would look like on each device similar to how we generate screenshots of what the drive looks like with files loaded for learn guide projects and library examples.
I'm not sure where would be the best place that could make use of them, but might be interesting if some guide or docs page for the various different devices could have an image depicting it for the specific device the page is talking about.
tulip sleet
I wonder if it would be helpful to make some code that parses the `pins.c` file ...
MicroPython uses a .csv file to specify the pins. It was harder to read, and there wasn't a way to add comments. But we could go back to a .csv or other easily parseable representation that had additional features we needed.
lime trellis
Question: does each microcontroller.nvm write cause an erase cycle for the entire NVM block (minimum still one memory page?). That's how I'm interpreting the docs
tulip sleet
Question: does each `microcontroller.nvm` write cause an erase cycle for the ent...
Theoretically it could notice that it's writing over one's and not bother to do an erase. You might check to see if it's doing that.
lone axle
I do think it has logic like that. I recall seeing it recently working on rp2040, don't know if it is the same on all devices though.
lime trellis
Theoretically it could notice that it's writing over one's and not bother to do ...
I'm more worried about repeated rewrites to all 8192 bytes. I use all sorts of bit flags that I store within a couple NVM bytes, but if manipulating these bytes is actually rewriting 8192 bytes then maybe it's not worth it?
tulip sleet
I'm more worried about repeated rewrites to all 8192 bytes. I use all sorts of b...
On SAMx5, the minimum write-size is a quad-word. You can't write individual bytes. The minimum erase size is 8k. We don't try to optimize quad-word writes to avoid erases.
If you are writing config data between deep sleeps, you could use alarm.SleepMemory, which does not use flash.
SAMx5 has a "SmartEEPROM" facility which we don't use, which does remapping to avoid erases. The nvm implementation could be changed to use that facility.
SAMD21 does not have that, but has an "EEPROM emulation" facility, which has a cache, and writes to flash on power fail as soon as possible (I think - I only looked briefly). The writing is handled by software, not hardware. http://ww1.microchip.com/downloads/en/DeviceDoc/Atmel-42125-SAM-EEPROM-Emulator-Service-EEPROM_ApplicationNote_AT03265.pdf
On nRF, in the bonding/pairing storage. I did try to avoid erases, and came up with a somewhat tricky data structure of valid and invalid data blocks, which are scanned in sequence to skip invalid entries
manic glacierBOT
lime trellis
On SAMx5, the minimum write-size is a quad-word. You can't write individual byte...
thank you for the reply! I'm digging thru the asf4 now trying to see where SAMx5 is doing the nvm writing/erasing.
I didn't implement alarm.SleepMemory when I wrote CP's samd alarm module and it looks like it remains that way.
EDIT: yep found it. it's in hpl_nvmctrl
lime trellis
@tulip sleet I'm now to the point where I'm seriously considering implementing this poorly documented "SmartEEPROM" for the samd. These two examples help a lot: https://community.atmel.com/comment/3037916#comment-3037916
https://community.atmel.com/forum/can-someone-explain-seeprom-sectors-me
.
.
but I could use some help preventing MP from using the smarteeprom address space. do you recall if we're using BANKA or BANKB of the NVM block?
manic glacierBOT
tulip sleet
<@!329766224093249548> I'm now to the point where I'm seriously considering impl...
We are using both banks. You would lay out the SmartEEPROM usage in the 'ports/atmel-samd/common.template.md' file. Another alternative is to use my nRF technique of writing small chunks, and invalidating the old chunks. I have a flag and length field at the beginning of each chunk. When a chunk is invalidated, the flag is set to 0. There's a linear search of blocks to find the first valid one. See ports/nrf/common-hal/_bleio/bonding.h and .c. There is a mix of specific (related to bonding) and general in there. The mechanism could be generalized.
Disadvantage of using SmartEEPROM is that it's SAMx5x-specific. Advantage is that it takes care of stuff for you (I don't know how well).
i looked in the errata but fortunately didn't see any cautions about the SmartEEPROM
tulip sleet
@onyx hinge @slender iron I am going out for a walk but I think I will make a beta.3 this morning, since the #5705/#5708 regression was major and is now fixed.
onyx hinge
@tulip sleet sounds good to me! .. just stopping in for a second, back to hiking soon. 🚶 🥾
tulip sleet
thanks for the time, enjoy yourself!
lone axle
@idle owl If you end up having a moment today I'm interested to get your thoughts on: https://github.com/adafruit/Adafruit_CircuitPython_PyPortal/pull/116 this PR would essentially revert one of recent changes to use the with context processor in this library. But it turns out that using that with context was causing an issue where it won't be able to play the wav file if wait_to_finish argument is set to False This issue has some more details about it: https://github.com/adafruit/Adafruit_CircuitPython_PyPortal/issues/118 Can also make a note for 'in the weeds' if you think it would benefit from more discussion.
idle owl
<@!330227457296957440> If you end up having a moment today I'm interested to get...
If it makes things work properly, then reverting it is fine. There was another lib for which we had to do the same. Or opposite? I don't remember now, but we've had to change use of with to make things work recently in other places too.
If that's the best way to do it, then roll with it.
lone axle
Okay, sounds good. I do recall a similar situation arising in a different library but not certain which one it was. Thank you.
idle owl
Okay, sounds good. I do recall a similar situation arising in a different librar...
It came out of the Dylan Pylint update sweep. She did the best she could with that, but it meant some things were done wrong.
Do you want me to merge the PR?
lone axle
Do you want me to merge the PR?
I can get it in a few minutes, but you can as well if you like.
idle owl
I can get it in a few minutes, but you can as well if you like.
I'll leave it to you then. Already onto going through email 🙂
@tulip sleet FYI, editing a proper username into a post doesn't tag anyone. It only tags on sending.
tulip sleet
ugh, thanks, I forgot that.
@lone axle those that can be changed to take a filename instead of an open file handle should do that. Does that make sense in the place you are mentioning?
lone axle
<@382939733107408897> those that can be changed to take a filename instead of a...
That thought occured to me as well, and I did fix a usage of OnDiskBitmap in PYOA library by making that change. But in PyPortal it's for a wave audio file not OnDiskBitmap. I don't think the core API allows the filename in that case.
lone axle
audiocore.WaveFile(wavfile) is the one having the issue. Though in the longer term it might be nice if that allowed a filename.
Although it may still need to have the option of using an open buffer like object as well. Because there are some RawSample examples that use that functionality I think to play sounds that don't originate from a file.
Oh, maybe those don't use WaveFile actually
I can take a look into it this week to see how OnDiskBitmap handles it and see if I can replicate it over into audiocore.WaveFile. If/When we do make that change in the core we'll be able to remove the pylint exception in PyPortal library which will be nice.
tulip sleet
feel free to open a tentative "Long Term" issue for that enhancement
there may even be one already
proven garnet
Question for anyone that knows displayio: Is this a bug, a lack of a feature, or something else:
stuck elbow
Question for anyone that knows displayio: Is this a bug, a lack of a feature, or...
it's there
lone axle
Question for anyone that knows displayio: Is this a bug, a lack of a feature, or...
My understanding is that the brightness property from the core displayio.Display class should work for that driver since it's extending displayio.Display https://circuitpython.readthedocs.io/en/latest/shared-bindings/displayio/index.html#displayio.Display.brightness
So if initializing the SSD1306 displayio driver and then attempting to set display.brightness raises an exception I would think that is a bug, I'm not sure if that is what is happening in this case though.
idle owl
@lone axle PyBadger wasn't working on 7.1.0 beta whatever. Simpletest complained about things. 7.0.0 works. Probably someone should look into it, I don't have time to. Could be a me-issue though.
lone axle
I can try it out on an Edge badge to see if anything shakes out on that device.
crimson ferry
strictly speaking, OLEDs don't have brightness, so a contrast control is used to simulate it
slender iron
right, but it's hooked up to brightness in displayio because it gives a similar effect
proven garnet
https://github.com/adafruit/circuitpython/blob/main/shared-module/displayio/Disp...
Thanks!
proven garnet
My understanding is that the `brightness` property from the core `displayio.Disp...
Awesome, thanks!
So it sounds like a bug, I can try and look into this
idle owl
<@&356864093652516868> Meeting in 30 minutes. If you're planning to participate or want your notes read off, please add your hug reports and status updates to the notes doc. Thanks and chat soon! https://docs.google.com/document/d/1Bwnt-pbFcrfo1gn_tH4L0HXkMXZbcqG5eoAOfPK1xco/edit#
formal summit
can circuitpython use SPI flash for storage? as the disk drive to run code from?
stuck elbow
@formal summit yes, that is what it does by default
formal summit
ok thanks, just wanted to confirm
because I'm currently running from the QSPI flash, and can't write to it
Are there any docs for using external memory?
tulip sleet
ok thanks, just wanted to confirm
Do you want to store files, or just bytes?
formal summit
I want to replace the fs
formal summit
that shows up as the USB disk
lone axle
I cannot run a backup recording today. I have some electrical work being done at the house. Would likely be lots of extra noise in the recording if I tried to do one.
idle owl
I cannot run a backup recording today. I have some electrical work being done at...
No worries, thanks for the heads up.
tulip sleet
I want to replace the fs
you need CIRCUITPY to hold the code. Is it that you can't write to CIRCUITPY from your code? You can remount CIRCUITPY as read/write instead of read-only
idle owl
brb!
formal summit
you need CIRCUITPY to hold the code. Is it that you can't write to CIRCUITPY fro...
I'm trying to build for a new stm32 board. By default the fs setup is happening on the internal flash
idle owl
@turbid radish You're unmuted.
tulip sleet
I'm trying to build for a new stm32 board. By default the fs setup is happening ...
have you looked at the other STM32 builds that use external flash: e.g. STM32F405 Feather
also I think there is a Discovery board with external flash
turbid radish
Please consider contributing
tulip sleet
@formal summit we are having an audio meeting at the moment, hence all the other posts
formal summit
have you looked at the other STM32 builds that use external flash: e.g. STM32F40...
I'll check this out first. not sure how external memory is used (I can only see supervisor_flash_write_blocks for internal memory)
hidden oxide
I guess I won't be missing some of the meeting after all
formal summit
<@371192862328750081> we are having an audio meeting at the moment, hence all th...
Sorry about the messages
tulip sleet
I'll check this out first. not sure how external memory is used (I can only see ...
see ports/stm32/supervisor/qspi_flash.c, and the code that calls what's in there
gilded cradle
no
turbid radish
Welcome Tim!
tulip sleet
you can read my notes
lone axle
Welcome Tim!
Thank you! So excited to be joining the team.
hidden oxide
Sorry, gotta go - just wanted to drop in and say hi
errant grail
Another hug to @lone axle for the NVM storage library. Already incorporated it into the espresso scale project to remember tare and alarm settings.
lone axle
Amazing photo from Jeff:
errant grail
Thank you!
gilded cradle
👋
ember iris
Thanks all!
blissful pollen
I have a new core addition for the IS31 work (that goes along with the python libraries that are in progress). The core PR is a pre-req to the python library. Should I be pushing this to the latest branch or? Just unsure with the beta releases going on
blissful pollen
thanks, it's a small bug fix and the equivalent of neopixel_write added to the module
solar whale
@slender iron ```Adafruit CircuitPython 7.2.0-alpha.0-104-g568c281fd on 2021-12-11; Raspberry Pi 4B with bcm2711
import os
os.listdir()
['.fseventsd', '.metadata_never_index', '.Trashes', 'code.py', 'lib', 'boot_out.txt']
f = open("boot_out.txt")
f.readline()
'Adafruit CircuitPython 7.2.0-alpha.0-104-g568c281fd on 2021-12-11; Raspberry Pi 4B with bcm2711\r\n'
🎉
solar whale
At first os.listdir did not show the boot_out.txt, but after a pwer cycle it appeared...
blissful pollen
So git question.. when I try to PR the 3 commits I have to the 7.1.x branch the PR tries to include a load of commits that I did not do. Anyone know if I'm missing something obvious? Do I have to somehow set my branch start to the 7.1.x start?
tulip sleet
e.g. in this non-CPy example, the "base:" dropdown on the left would be set to 7.1.x:
blissful pollen
I switched the base to 7.1.x but then I see "commits 58" most are not mine of course
solar whale
hmm - setback. still booting ok but now os.listdir() returns an empty list ....
tulip sleet
so was your PR branch not started from 7.1.x?
solar whale
even after earse_filesystem()
blissful pollen
Probably not as I started on the branch before 7.1.x
(sorry git newbie still - as much as I think I keep learning more 🙂 )
do i rebase my branch to the 7.1.x and then PR that?
slender iron
that's what I usually do
slender iron
even after earse_filesystem()
I'm busy at the moment but can look more later. If listdir is empty, that seems like the SD card didn't start at all
blissful pollen
okay thanks I think I can figure that out
slender iron
Here is the notes document for Monday’s CircuitPython Weekly meeting. It is at the normal time of 11am Pacific / 2pm Eastern here on Discord. Everyone is encouraged to attend! Please add your hug reports and status updates even if you’ll be attending the meeting - it’s super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and we’ll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/1L-I0fOE6yrnv7OV_jQZZiFGtgWmc86GiGP6_ecpnGQs/edit?usp=sharing
solar whale
I'm busy at the moment but can look more later. If listdir is empty, that seems ...
No rush -- I can't do much more with it today. Some successes. Current status. It does boot OK from a USB drive -- So far I don't see any file system though. Reflashing and booting from SDCard does have a file system...for awhile, then it seems to go away...
digital shoreBOT
adafruit
blissful pollen
Sorry lost in git again. I'm missing something to rebase my branch onto the latest 7.1.x branch. I tried git rebase adafruit/7.1.x but the parent commit is still a few days off the latest 7.1.x branch like it rebased to the start of 7.1.x
crimson ferry
my build environment is completely messed up since esp-idf 4.4, I'll probably burn it all to the ground and start clean
lunar gull
my build environment is completely messed up since esp-idf 4.4, I'll probably bu...
TRIGGER WARNING! Man, I can relate... trying to build Apollo3 with Artemis stuff with esp-idf and whathaveyou. I think I have PTSD.
slender iron
Sorry lost in git again. I'm missing something to rebase my branch onto the late...
Did you fetch before doing the rebase? maybe you had an old version of the branch
sharp jacinth
Are some circuitpython libraries likely to be re-written now that asyncio is becoming a thing?
blissful pollen
Did you `fetch` before doing the rebase? maybe you had an old version of the bra...
Yup for the git fetch adafruit first. I’ll play more again later.
proven garnet
Catching up on the meeting today, congrats @lone axle!
manic glacierBOT
analog bridge
@tulip sleet Thanks for your suggestions on #5422, I think it should be good to go now.
tulip sleet
<@!329766224093249548> Thanks for your suggestions on #5422, I think it should b...
Thanks for the reminder. Too far down on my stack.
proven garnet
Oh @tulip sleet just wanted to say that asyncio is working great on that self-lighting menorah project. It seems that connecting to the Wi-Fi network is synchronous but that's not a huge issue and makes sense, but I love that connecting to the NTP server (which is the bulk of the time honestly) works very well asynchronously. If you want to look at the implementation, here it is, but if not then just a general thanks again for all of this! https://github.com/tekktrik/auto-menorah
Also, the guide you wrote helped a lot, I meant to mention that, especially how to keep track of a "status-like" object instance between coroutines
tulip sleet
Oh <@!329766224093249548> just wanted to say that asyncio is working great on th...
Glad the guide was helpful! If you wrote setup_menorah() as non-async, would it get stuck somewhere? I was thinking the actual lighting and checking was done with two tasks, not just the setup.
proven garnet
Glad the guide was helpful! If you wrote `setup_menorah()` as non-async, would i...
No it wouldn't, it would still work. The major benefit is that setup_menorah() allows me to manipulate the "candes" with display_loading() while I run setup_connections() which connects to Wi-Fi then NTP server for time. This is because the latter takes some time so manipulating the lights lets the user know that the menorah is alive and actively working on things and not just kaput haha
After all that's done, I move into main() to figure out which candles to light
tulip sleet
Ahl, got it, I didn't know that was your goal. It's a "Loading" status bar 🙂
proven garnet
Ahl, got it, I didn't know that was your goal. It's a "Loading" status bar 🙂
Yup!
tulip sleet
Honestly, I don't have much experience writing async task programs, but I tried to come up with the cleanest coding styles I could think of. A lot of the examples I have seen are distressingly confusing (like bad threading code)
proven garnet
I think you succeeded; I had never written async task programs (only limited knowledge of the concept too) and was able to get what I needed up and running!
lone axle
Is there a CircuitPython driver for this eeprom breakout? https://www.adafruit.com/product/5146
lone axle
I see the listing page links to an Arduino FRAM/EEPROM library. I know there is a CircuitPython FRAM library, but it doesn't seem to have any references to this EEPROM breakout in it.
If there isn't currently a CircuitPython driver (and no one is already working on it) I may make an attempt at creating one. Would be my first time working on lower level I2C or SPI communication so not sure if I'll get it working.
Would it make sense to add that to the existing CircuitPython FRAM library since the Arduino library does seem to bundle them together? Or should it be a completely separate library?
jaunty juniper
Is there a CircuitPython driver for this eeprom breakout? https://www.adafruit.c...
ah I had the same question, it doesn't have a learn guide either
lone axle
LadyAda mentioned a CircuitPython driving coming for it in the video introducing it I think. My guess is it's not created yet. But I also don't have any experience with FRAM or EEPROM breakouts. Maybe there is some chance that the existing FRAM library does already work with it.
jaunty juniper
yeah I was thinking of looking at the FRAM library too
formal summit
Hey @tulip sleet , thanks for your help. filesystem_init isn't failing anymore
tulip sleet
excellent
formal summit
The whole SPI flash selection was surprisingly easy. I'm a big fan of the code
tulip sleet
thanks! we had to support so many chips, we really needed to parameterize it
tannewt did nearly all of that
formal summit
It's magic 😄
lone axle
Was there a newer revision of the MagTag silkscreen that labels the buttons with letters A thru D? I noticed on board that there are pins labeled that way:
'BUTTON_A', 'BUTTON_B', 'BUTTON_C', 'BUTTON_D',
But on the silkscreen of mine and all of the images I can find I don't see any label letters. They have labels with the IO pin # i.e. D15, D14 etc.
proven garnet
I see the listing page links to an Arduino FRAM/EEPROM library. I know there is ...
I totally didn't order 10 eeprom boards before checking whether a library existed 
So this makes me happy
tulip sleet
Was there a newer revision of the MagTag silkscreen that labels the buttons with...
I don't think there is any silk with those button labels. The https://learn.adafruit.com/adafruit-magtag/pinouts#buttons-3077173-27 section of the guide should probably list the correspondences.
the prettypins diagram does have (because it's generated partly from pins.c)
lunar gull
Was there a newer revision of the MagTag silkscreen that labels the buttons with...
mine have little arrows and pin labels, but no A-D
and if I remember correctly they're left to right, whereas the D pins are reversed, and they skip D13 (for obvious reasons, can't go inviting bad luck like that 😉 )
ember iris
Was there a newer revision of the MagTag silkscreen that labels the buttons with...
I have one with a white face, and one with a black face. The white face was first, There's a couple of differences, let me take a pic
lone axle
board does also have properties for them as labeled on the silkscreen so folks could still have code match the silkscreen. I do like the existence of button_A, button_B etc.. as well because many of the other devices with multiple buttons do use those like pybadge and CPX/CPB so I could see folks naturally assuming they exist like that.
lone axle
and if I remember correctly they're left to right, whereas the D pins are revers...
yep this looks correct. button A has the highest D# value and they do skip D13.
lunar gull
if I remember, D13 is mapped to something else, as is standard on a lot of boards... ugh
lone axle
I will make a note to update the pinouts page buttons section to add the mapping there as well so it's an extra place people can find it. Really neat that pretty pins pulls from pins.c so they should always match.
lone axle
Yeah D13 is often an on-board LED for lots of devices.
ember iris
lone axle
Thank you!
ember iris
Neopixels are labeled as #, Neopix is a different digital pin, buttons are labeled identically
Oh I've got a bit of tape over the 'ON' LED on the black one--that LED is there
I can also try any script you need on the two magtags
lone axle
I can also try any script you need on the two magtags
Thank you! If you're up for it, would be much appreciated if you want to test / review this PR: https://github.com/adafruit/Adafruit_CircuitPython_PyBadger/pull/56 there is a magtag simpletest script in the examples dir that can be used to test it.
tulip sleet
i was just starting to review it when you changed it out from under me 🙂
lone axle
(testing on either one of them should be suffecient, no need for both.)
tulip sleet
@ember iris if you can test that would be great. I was not going to test on real hw
lone axle
Whoops sorry about that. I ran pre-commit locally the first time but didn't notice until too late that sphinx failed. new commit gets it back to happy I think.
lone axle
I wonder if we could fold the sphinx build into pre-commit (or even if we would want to if so). I seem to forget one or the other sometimes.
ember iris
Thank you! If you're up for it, would be much appreciated if you want to test / ...
(going slow so I can learn git remote checkout things but I am working on it)
lone axle
(going slow so I can learn git remote checkout things but I am working on it)
Thank you if you like CLI tools and haven't looked into it yet I have become a big fan of the Github CLI tool https://cli.github.com/
and no worries about pace.
lone axle
I think circup might be having trouble installing libraries ATM. Possibly something related to removing support for 6.X bundle recently?
❯ circup install adafruit_io
Found device at /media/timc/CIRCUITPY, running CircuitPython 7.1.0-beta.3-86-gf6de956c5-dirty.
Downloading latest version for adafruit/Adafruit_CircuitPython_Bundle.
[####################################] 100%
There was a problem downloading the bundle. Please try again in a moment.
I manually deleted the 6.x bundle that was built last night, because it contained a 6.x-incompatible library, and I had merged the "remove 6.x" PR this morning.
but I don't know what it is trying to download
lone axle
Maybe it's stuck trying to fetch the deleted one and it will go back to normal / notice that it doesn't need to tomorrow. I'll see if I can narrow it down to the specific part within circup rasing the error.
tulip sleet
i am going for a walk but will be back in an hour or so or less
lone axle
I'll open an issue on circup repo as well and we can see if other folks are noticing it. I've only tried a few times today so it could possibly be something on my end as well.
tulip sleet
I'll open an issue on circup repo as well and we can see if other folks are noti...
do circup --verbose ... and it will show it is indeed trying to download the 6.x bundle
lone axle
ooh nice didn't know about that flag. Thank you.
ember iris
and no worries about pace.
Ok I've tested it on both magtags, and the test worked identically and correctly on both. I did hit some odd errors, but they don't seem to be pybadger related.
Since they're not really pybadger related, I figured I'd bring them up here not on the pull request unless you think otherwise is best
main furnace
12/14/2021 12:42:59 WARNING: Unable to connect to https://github.com/adafruit/Adafruit_CircuitPython_Bundle/releases/download/20211214/adafruit-circuitpython-bundle-6.x-mpy-20211214.zip
There was a problem downloading the bundle. Please try again in a moment.
jaunty juniper
ooh nice didn't know about that flag. Thank you.
circup also outputs to a log file
I believe it's just a matter of updating the PLATFORMS config variable
lone axle
Since they're not really pybadger related, I figured I'd bring them up here not ...
Yep that sounds good to me. Can start the discussion here and then open a new issue on whichever repo would be most appropriate if needed.
lone axle
circup also outputs to a log file
Nice, I didn't know that either. Thanks for the tip.
ember iris
Yep that sounds good to me. Can start the discussion here and then open a new is...
Sounds good, Let me approve the pull request first in that case
lone axle
And yep removing the 6s from PLATFORMS does fix. Got a PR created for it here: https://github.com/adafruit/circup/pull/132
jaunty juniper
maybe in the future that could be something that is retrieved from the bundle repo or something
ember iris
Yep that sounds good to me. Can start the discussion here and then open a new is...
When transferring code, (my old code to pybadge, and pybadge to my old code) I'd end up with OSError: [Errno 5] Input/output error. To get this I'd delete everything and copy/paste the new code in. It would hit the error, then try again, and hit the error every couple of seconds. To get out of the loop I'd have to open code.py and save a change. I imagine this is standard circuit python behavior when you delete the code.py file and copy a new one in, but I hadn't run into it.
The other error was an invalid bmp format, the first time I tried to display the image card. I tried again and it worked. that only happened once on one device so it's probably more likely that something like a code reload happened and messed it up.
lone axle
I am unsure about the Input/Output error. With regards to the bmp card though I can think of 2 things that may be relavent:
- If you pasted every thing "at the same time" the OS may have actually pasted them sequentially. If code.py went first it would cause a reset and the bmp file might not be finished pasting completely yet so it could think it's an invalid bitmap if it tries showing it before it has the complete file.
- the image badge is the one type where the
display.refresh()call happens internally in the library. On EInk displays if you call refresh() too soon after a previous refresh it will raise an exception (though the text in it should be more specifically about refreshing too soon). The magtag example script works around this for the other types of badges by implementing it's on try/except logic.
We should probably either make the library use the same refresh behavior with all types and adjust the example as necessary to reflect that. OR maybe just add the slightly different try/except logic for the bmp type badge in the magtag example.
ember iris
I think 1 is probably what occurred, it was after a copy paste. I tried to stress it with the refreshing too soon and it handled that portion gracefully
idle wharf
Are there any repos in the Adafruit managed repos or CP_org that create PRs from GH Actions? There's a host of ways, and I wanted to see if there's some way its already being done as I work to wrap up show-note automation for Deep Dive repo.
manic glacierBOT
I do not think so. I tried to format one SD card with RPi Imager and put second one in to the RPi, which was previously also flashed via RPi Imager. I checked what partiton it has before insert and It was FAT32, which is same as at this card with letter F.
idle wharf
Thank you @tulip sleet I'll take a look at that one!
tulip sleet
it updates circuitpython.org with new versions and new boards
when there is a new release
proven garnet
Is there any value in doing typing for things not already tagged with the associated issue? Was curious about PlatformDetect and noticed it didn't have typing or the issue.
tulip sleet
Is there any value in doing typing for things not already tagged with the associ...
PlatformDetect is not typically used directly by end users: they don't bring it into their editors, etc. So adding annotation is a lower priority. But I don't think it would be bad. You could add it as an issue and ask in the issue whether it would be helpful.
proven garnet
PlatformDetect is not typically used directly by end users: they don't bring it ...
Sounds good to me!
tulip sleet
@daring pumice I'd be happy to have audio conversation about the pullups issue and the PR during the day on Wednesday. We might save some time that way, both to discuss the general issue and the technical details.
vague thicket
@slender iron what approach would you recommend please, if I would like to try make touchscreen working with ads7846 driver IC and RPi? I tried it with adafruit_touchscreen, but it does complain that there is no analogio module. Thanks
slender iron
sounds like touchscreen should be reworked to pass in the analogin object
that way you could pass in something that does ADC over a bus rather than just native
vague thicket
by touchscreen you mean physical touchscreen or the library?
slender iron
the library
vague thicket
is this right place to look what to bend? https://github.com/adafruit/Adafruit_CircuitPython_Touchscreen/blob/main/adafruit_touchscreen.py
slender iron
yup, that's the right spot
looks a bit trickier than I expected due to the digitalio stuff
jaunty juniper
I have seen a few people being confused because they couldn't find "boot.py", not aware that it's just a file they can create, I wonder if there is some wording in some guide(s) that could be adjusted to clarify that or if it's just some random chance.
jaunty juniper
like the customizing USB devices guide says "you can write code in the boot.py file", or the code "must be put in boot.py". It seems obvious to me that you can create it if absent, but to some people it might sound like telling them to edit some existing file they can't find ? (Like editing some ini file on windows or something)
viscid pine
i see how that could be confusing if they are used to there just being one .py file
meager adder
The latest 20211215 bundle is missing the zipped assets: https://github.com/adafruit/Adafruit_CircuitPython_Bundle/releases/tag/20211215
vscode-circuitpython (https://github.com/joedevivo/vscode-circuitpython) detects an update but there is no zipped asset to download so it gets stuck and nothing works, not even the serial monitor.
Is this normal and the zipped assets just gets uploaded at a later time?
crimson ferry
@meager adder It just takes a little time
sometimes it does need some manual nudging... https://github.com/adafruit/Adafruit_CircuitPython_Bundle/actions
meager adder
sometimes it does need some manual nudging... <https://github.com/adafruit/Adafr...
It looks like upload-release-assets failed. I guess this means it will never upload the assets until someone rolls back or tries to run it again?
Building mpy-cross for circuitpython 6.2.0
==========================================
Traceback (most recent call last):
File "/opt/hostedtoolcache/Python/3.6.15/x64/bin/circuitpython-build-bundles", line 11, in <module>
load_entry_point('circuitpython-build-tools==1.8.5', 'console_scripts', 'circuitpython-build-bundles')()
File "/opt/hostedtoolcache/Python/3.6.15/x64/lib/python3.6/site-packages/click/core.py", line 1128, in __call__
return self.main(*args, **kwargs)
File "/opt/hostedtoolcache/Python/3.6.15/x64/lib/python3.6/site-packages/click/core.py", line 1053, in main
rv = self.invoke(ctx)
File "/opt/hostedtoolcache/Python/3.6.15/x64/lib/python3.6/site-packages/click/core.py", line 1395, in invoke
return ctx.invoke(self.callback, **ctx.params)
File "/opt/hostedtoolcache/Python/3.6.15/x64/lib/python3.6/site-packages/click/core.py", line 754, in invoke
return __callback(*args, **kwargs)
File "/opt/hostedtoolcache/Python/3.6.15/x64/lib/python3.6/site-packages/circuitpython_build_tools/scripts/build_bundles.py", line 264, in build_bundles
mpy_cross=mpy_cross, build_tools_version=build_tools_version)
File "/opt/hostedtoolcache/Python/3.6.15/x64/lib/python3.6/site-packages/circuitpython_build_tools/scripts/build_bundles.py", line 154, in build_bundle
mpy_cross=mpy_cross, example_bundle=example_bundle)
File "/opt/hostedtoolcache/Python/3.6.15/x64/lib/python3.6/site-packages/circuitpython_build_tools/build.py", line 257, in library
raise RuntimeError("mpy-cross failed on", full_path)
RuntimeError: ('mpy-cross failed on', '/home/runner/work/Adafruit_CircuitPython_Bundle/Adafruit_CircuitPython_Bundle/libraries/helpers/asyncio/asyncio/funcs.py')
Error: Process completed with exit code 1.
I suspect it will get addressed in the morning, or deferred to the next evening for a good run
depending on what's needed to fix
looks like maybe the prior day manual asyncio fix is needed again?
meager adder
Ah ok. I'm kind of stuck right now because vscode-circuitpython refuses to do anything (not even serial monitor) until the new bundle is downloaded, but there is no new bundle to download. But I guess that is also more an issue with the extension and not with circuitpython specifically
crimson ferry
this is kind of unusual, but I wonder if there's a way to file an issue with vscode
meager adder
The issue will probably be gone by the time someone looks at it
Is it possible to have the release pipeline start with a private release and only promote it to public iff all steps in the pipeline pass (including asset upload)?
There are probably other external processes that rely on bundle updates and rather than fixing them individually by every consumer, addressing this upstream in the circuitpython pipeline might be easier
jaunty juniper
hmmm why does it say Building mpy-cross for circuitpython 6.2.0 ?
main furnace
pi zero 2w: Can't read from SD card
Adafruit CircuitPython 7.2.0-alpha.0-127-gcad8c1fcd on 2021-12-15; Raspberry Pi Zero 2W with bcm2837
I'm booting untethered (using the serial console).
Works if I revert to alpha.0-50-g8eee16c7a 12-08 build.
read sectors result 3 when reading block 531865 for 1
tulip sleet
@meager adder @crimson ferry I am working on the bundle build failure. It was still trying to build 6.x stuff.
solar whale
`read sectors result 3 when reading block 531865 for 1`
Whew! I though it was only me -- seeing the same
looks like the error comes from here https://github.com/adafruit/circuitpython/blob/main/ports/broadcom/common-hal/sdioio/SDCard.c#L334
manic glacierBOT
tidal kiln
anyone here on MacOS and using a board with a CP2104 USB-to-TTL chip?
solar whale
anyone here on MacOS and using a board with a CP2104 USB-to-TTL chip?
I'll try -- what board are you using?
tidal kiln
thanks. doesn't matter. just wanting to know what it shows up as in /dev
so any CP2104 board should be fine
tidal kiln
hmm. it creates two entries?
solar whale
yes -- both work -- to get to REPL on MicroPyhton anyway.
I "think" it is a CP2104
yes -- chip says SIL 2104
tidal kiln
feather esp32?
solar whale
yes
tidal kiln
yep. it has one.
solar whale
any other info/tests you need?
tidal kiln
@solar whale nope. thanks. not unless you know of reasons only the SLAB entry would show up.
solar whale
<@!343570430042308609> nope. thanks. not unless you know of reasons only the SLA...
No, sorry -- it's all magic to me....
proven garnet
@gilded cradle I'm adding licenses to PlatformDetect, should I add your name to them or is there someone better. Doing as part of getting pre-commit running.
slender iron
@crude blaze do you have a USB PID for piunora? Are you planning to open source the design? I started making a board def for it in CP.
crude blaze
def. yes on open source design, I just have to find the time to clean up the Eagle files
will apply for oshwa cert then
I do not have PID for Piunora. I would still need one for CPY
slender iron
right, you can get a pid.codes one once you post the hardware files
crude blaze
right
slender iron
I have it using a test pid at the moment
crude blaze
I'll try to get around to it soon
crude blaze
schematic is here already if that helps
https://cdn.hackaday.io/files/1784987636858432/Piunora.pdf
slender iron
@crude blaze my branch is here: https://github.com/tannewt/circuitpython/tree/piunora
crude blaze
awesome, thank you for doing this. I'll try to check it out this week
main furnace
Git help? I did checkout -b mybranch origin/7.1.x, now I have a heap of 'changes not staged for commit'. I just want to fix a typo in one file, this doesn't look right.
main furnace
Everything is frozen|lib|ports|tools|data|extmod.
Fetch-submodules cleaned up most of it.
main furnace
Plan B: edit files on GitHub.
tulip sleet
Plan B: edit files on GitHub.
git checkout 7.1.x
git fetch adafruit
git merge adadfruit/7.1.x
will get your fork up to date
then
git checkout -b name-of-pr-branch
you would already have done:
git remote add adafruit https://github.com/adafruit/circuitpython
My process for a new change is:
git fetch adafruit
git checkout adafruit/7.1.x
git switch -c my_branch_name
git status
make fetch-submodules
git status
daring pumice
Evening, everyone 🙂
main furnace
Guess I need to update my build tools. Uncrustify 0.68_f can't do it.
onyx hinge
A job matrix can generate a maximum of 256 jobs per workflow run. This limit also applies to self-hosted runners.
We may hit this limit sometime in the next year, as all our "arm" builds are from one "matrix".
onyx hinge
testing this morning with an up-to-the-second circup [with removed 6.x bundle support] & it works again. yay.
jaunty juniper
are you making a (circup) release then ?
onyx hinge
https://github.com/adafruit/circup/pull/129 I'm looking at this PR and wondering .. does circup actually use babel?
oh it's something to do with rtd https://python-project.readthedocs.io/en/latest/simple.html#creating-a-new-project
As requirements file, enter rtd_requirements.txt. RTD will pip install -r rtd_requirements.txt. The rtd_requirements.txt of the template references the optional dev dependencies, so you can forget about this file and add doc dependencies to extras_require['dev'] in setup.py.
@jaunty juniper https://github.com/adafruit/circup/releases/tag/1.0.1
should be available on pypi within the hour 🙂
jaunty juniper
thanks !
tulip sleet
@analog bridge Re turning off RGBMatrix/Protomatter due to ESP-IDF 4.4: Is this fixed by https://github.com/adafruit/Adafruit_Protomatter/pull/41 ? Is there an upstream fix or issue in esp-idf? I could not find such a thing. THanks.
onyx hinge
@tulip sleet I think one element of this issue is that esp-idf renamed registers. If you turn RGBMATRIX on and build for kaluga (previously supported), you get ```In file included from ../../lib/protomatter/src/arch/arch.h:168,
from ../../lib/protomatter/src/core.c:33:
../../lib/protomatter/src/arch/esp32.h: In function '_PM_timerGetCount':
../../lib/protomatter/src/arch/esp32.h:192:42: error: 'timg_txupdate_reg_t' {aka 'volatile union <anonymous>'} has no member named 'update'; did you mean 'tx_update'?
timer->hw->hw_timer[timer->idx].update.update = 1;
^~~~~~
tx_update
../../lib/protomatter/src/arch/esp32.h:193:41: error: 'timg_hwtimer_reg_t' {aka 'volatile struct <anonymous>'} has no member named 'cnt_low'
return timer->hw->hw_timer[timer->idx].cnt_low;
^
../../lib/protomatter/src/arch/esp32.h:194:1: error: control reaches end of non-void function [-Werror=return-type]
}
^
cc1: all warnings being treated as errors
yup, the PR just seems to change some names.
onyx hinge
looks like the related header just got fully rewritten in esp-idf, with incompatible names. so the esp32s2 part should be perfectly safe to take. we can leave RGBMATRIX disabled on s3 & c3 until someone actually tests it on a panel.
solar whale
FYi -- from mouser --- looks like the SAOLA is nearing end of life --- time to move on....it was worth the $8 (now 9)
