#circuitpython-dev
manic glacierBOT
onyx hinge
sadly, even with the 1500 bytes I found, and another 170 bytes, I still can't enable built-in sdcardio on feather m0 adalogger. π
I could disable neopixel too (there's no onboard neopixel)
solar whale
@onyx hinge I'll be happy to create a PR for the BlackPill. I have not done one since the "pre-commit" changes were implemented. Can you easily describe what I need to do now to create a PR -- I don't see any mention of "pre-commit" in the guides for PR ... or have I missed it?
onyx hinge
@solar whale I think someone was updating a guide to add it, not sure
blissful pollen
I'd think you could probably get away without neopixel for a logger, or even put a note that if you want this get X build instead?
onyx hinge
I don't see where it was added to docs. you want to pip install pre-commit (however you invoke pip normally, it may be pip3, python3 -mpip, etc,) once per system. Then in any git repo where you want to enable pre-commit checks, you run pre-commit install once per repo to enable checking each time you commit and you can check all files with pre-commit run --all.
@solar whale ^
solar whale
@onyx hinge Thanks -- I'll give it a try later today.
onyx hinge
It would be nice to have sdcardio built in, but back over 1600 bytes of flash storage savings right away -- when there's a perfectly workable if not quite as fast module to do the same thing -- seems like a wrong decision.
tuning up performance a little bit on an M0 board is less important than just giving us some everyday breathing room when it comes to M0 builds filling their flash
especially since people would have to change their code to use it
solar whale
That is a strange case where you run out of Flash, but have enough RAM to load the Python module....
onyx hinge
Β―_(γ)_/Β―
manic glacierBOT
tulip sleet
@onyx hinge I'm fine with merging the COMPUTED_GOTO savings. I left it to you in case you had more to add.
thorny jay
What did go wrong with my PR #32 to split this library into package to save some space? It was an answer to an issue #27 from @slender iron from April 7th 2020 and my PR was available since August 24 2020. Now around February 14 2021 Sandy made a PR for new hardware from Pimoroni (and it is great that RP2040 based Pimoroni hardware use CircuitPython). But now my PR cannot be reconciled automatically (or at least you may understand that I don't want to spend time on this again...). https://github.com/adafruit/Adafruit_CircuitPython_IS31FL3731/network
tulip sleet
What did go wrong with my PR #32 to split this library into package to save some...
Hmm, usually GitHub is better at showing the merge conflicts. I think this was held up only because it requires changing a bunch of existing code and examples. Feel free to add a comment to the PR that you don't have time to work on this at the moment.
thorny jay
Hmm, usually GitHub is better at showing the merge conflicts. I think this was h...
Yeah, it did require change in learn guide, those were identified. I believe it is still needed to split as this library will continue to inflate with stuff like this: https://twitter.com/veryalien/status/1368573430656864257?s=20
And the day such a change will be done, it will affect all of the Pimoroni customer too... because they started on the non split library.
thorny jay
Hmm, usually GitHub is better at showing the merge conflicts. I think this was h...
Maybe @slender iron was right and each product should have it's library and not sit in the chip library: https://github.com/adafruit/Adafruit_CircuitPython_IS31FL3731/issues/27
tulip sleet
yes, that sounds like how displayio additions libraries vs drivers is structured
thorny jay
But this is a core work for Adafruit, not from contributor.
thorny jay
If Adafruit show the way, contributor will follow the pattern. When I added board to that library, I followed the existing pattern.
tulip sleet
If Adafruit show the way, contributor will follow the pattern. When I added boar...
I think this is worth discussing in https://github.com/adafruit/Adafruit_CircuitPython_IS31FL3731/issues/27
manic glacierBOT
Thank you @hathach and @dhalbert.
I can confirm that setting serial.DtrEnable = true; in C# solved the issue:
Would be amazing if this was documented, to avoid possible duplicated issues.
thorny jay
I think this is worth discussing in <https://github.com/adafruit/Adafruit_Circui...
Done
stone iris
Seeing reports of Mu not recognising the REPL when using a Pico with the latest beta of CircuitPython.
blissful pollen
Seeing reports of Mu not recognising the REPL when using a Pico with the latest ...
I had it occur to me where the second USB CDC showed up as the first COM port (ie. COM3 and the REPL on COM4). A reboot solved it so I thought it was an issue on my end, but maybe a race which COM gets numbered first?
turbid radish
?serverinfo
digital shoreBOT
adafruit
lone axle
Are there any other relevant flags similar to --all-files that can tell pre-commit to run only on checked in files?
--files [FILES [FILES ...]] seems like it could help but I'm not sure if all the hooks treat that parameter the same. I tried using that instead of all-files but some of the hooks still failed for some things unrelated to the files I specified, so some of those hooks must be looking at a wider set than the ones specified by this.
tulip sleet
Are there any other relevant flags similar to `--all-files` that can tell pre-co...
not that i can see. you can git stash your uncommited changes to hide them and then git stash pop to get them back. Some peole have more complicated ways: https://stackoverflow.com/questions/13780700/how-can-i-run-git-pre-commit-checks-only-on-staged-content
analog bridge
Are there any other relevant flags similar to `--all-files` that can tell pre-co...
git diff --cached --name-only | xargs pre-commit run --files might work.
lone axle
Thank you that does seem to make it skip my "extra" uncommitted file (perhaps using the stash technique mentioned above) But it also seems to have caused it to skip an actual change in one of the real library files:
Actual output did not format well. It shows skipped on all of the checks including pylint and black, which I think it needs to run on my change.
maybe the git diff needs to be done without --cached removing that does get it to list the file I changed.
fossil gorge
I thought that pre-commit automatically stashed anything that wasn't ignored or staged in Git. Didn't it?
lone axle
It does print these messages which seem to indicate it is:
[INFO] Stashing unstaged files to /home/timc/.cache/pre-commit/patch1615213587.
....running checks...
[INFO] Restored changes from /home/timc/.cache/pre-commit/patch1615213587.
But it's also still failing with reasons pointing to files that aren't checked in. I'm not sure I understand the behavior of all the different components involved.
I don't get those messages every time I run either though. Perhaps that is tied to one of the input flags.
tulip sleet
dang that was pressing the wrong button [accidental beta.4 release - removed]
orchid basinBOT
onyx hinge
oops!
tulip sleet
I was trying to do "Save Draft" in the release notes, and 2 pixels over was "Publish Release". oh groan, well, I could do it, but it's not that interesting. I'll get Anne to kill the tweet.
manic glacierBOT
Now, reset cpu after wake up from pretending deep sleep. The same behavior between pretending and native deep sleep.
manic glacierBOT
I would expect somewhat lower, based on the chart above (clipped to System ON states below). But it could be a lot of things. Did you measure in System OFF also? There may be further things to adjust, but this is quite good for now, and I'll turn off Draft status for the PR.
manic glacierBOT
thorny jay
https://github.com/adafruit/Adafruit_CircuitPython_IS31FL3731/blob/a85d01575b0e95530570d9e48289b34cd47fedc7/adafruit_is31fl3731/matrix_11x7.py#L45 Wondering how much is too much optimisation vs readability as this one liner does the same for all usefull value of x and y: return (x<<4) + y - (0 if x<=5 else 88)
versus
if x <= 5:
r = x * 16 + y
else:
r = (x - 5) * 16 + y - 8
return r
tulip sleet
if you put the original formulas in a comment, I think that's fine. Then someone can check your algebra and recognize the original formula from the datasheet or whatever. If the rewrite reduces the number of computational operations, then it seems worth it.
comments are free and save the reader time
drowsy geyser
FWIW I think the one-liner is more elegant, but as @tulip sleet says, a comment would be justified.
thorny jay
I'll check some of those formula, the original was a truth table, I made the if else, then I found about tertianary operator.
I have a test program to compare two function and make sure they give the same result. π
tulip sleet
i have done some minor timing checks on tertiary operator vs if's and I don't remember that there was much difference.
drowsy geyser
Interesting. I'm trying to build latest from source, but I'm getting an error about strict overflow checking: CC ../../lib/tinyusb/src/common/tusb_fifo.c ../../lib/tinyusb/src/common/tusb_fifo.c: In function 'tu_fifo_write': ../../lib/tinyusb/src/common/tusb_fifo.c:149:6: error: assuming signed overflow does not occur when assuming that (X + c) < X is always false [-Werror=strict-overflow] if ((p > p + offset) || (p + offset > f->max_pointer_idx)) ^ cc1: all warnings being treated as errors make: *** [../../py/mkrules.mk:55: build-feather_m4_express/lib/tinyusb/src/common/tusb_fifo.o] Error 1 Could someone point me to the right solution (I don't want to just bodge it)?
tulip sleet
make sure you are using gcc10
check the Building CIrcuitPython guide
gcc 10 is smarter about that error
drowsy geyser
Hum. No joy with gcc-10. I'll just pull latest from the S3 bucket.
modern wing
Good afternoon all you wonderful folks! 
idle owl
@gilded cradle Do I see correctly that you submitted the tests for the datetime library?
ionic elk
Won't be able to make today's meeting unfortunately
gilded cradle
I was mostly disabling skips
ionic elk
Didn't really have anything new anyway though
gilded cradle
Each file does multiple tests
idle owl
@gilded cradle Ok. Pylint is having a rough time with it, and I think it's worth discussing how we run Pylint on tests. But I also think some of it might be valid failures. I'll bring it up in the meeting.
idle owl
Each file does multiple tests
Oh I get what you mean.
turbid radish
Lurking
drowsy geyser
Also lurking. But hello everyone.
modern wing
better!
solar whale
@drowsy geyser did not recognize the name change...
gilded cradle
Same here, but I recognized the profile pic
modern wing
@idle owl Macbook?
thorny jay
array of microphone.
solar whale
Magic Mic
idle owl
@modern wing MacBook Pro. yes.
drowsy geyser
@solar whale It's easier to pronounce than "hukuzatuna"
modern wing
@idle owl It's probably under the left side speaker -- on the 15" and 16" models, it mutes the left side speaker to help with the echo cancellation. I've found out the hard way when touching the left speaker was....unkind....to my listener's ears. π
idle owl
@modern wing Gah yeah. Fair enough.
FYI: The US changes to Daylight Saving Time on March 14, so the next meeting is at 2PM EDT, which is UTC-4.
slender iron
hey folks and <@&356864093652516868> , feather rp2040s are in stock if you want one: https://www.adafruit.com/product/4884
idle owl
@inland tusk Please mute.
solar whale
wrong channel
thorny jay
not one but all the answer. π
That game beat me every time... I used to be good.
turbid radish
modern wing
Good luck with the new one -- we had the same happen a few years ago, and disliked the new one so much we returned it and got a used Speed Queen.
thorny jay
Is there a CP chess yet?
turbid radish
Maybe sunfish could be used?
manic glacierBOT
solar whale
is the beta.4- tag gone too?
slender iron
we should have the next one be beta.5. it's ok to burn beta.4
turbid radish
Email me @lone axle when the guide is submitted please
tulip sleet
is the beta.4- tag gone too?
yes, the tag is deleted
solar whale
You need to make a .mpy to test on an M0 board
still zephyr
Talking about capturing folks https://www.youtube.com/watch?v=07vG-_CcDG0&t=1153s
thorny jay
Now I am worried because I don't know if I have been using VSYS or VBUS...
idle owl
gilded cradle
I'm pretty sure that's correct.
fossil gorge
Here but lurking
turbid radish
Thanks everyone
modern wing
Thanks π
gilded cradle
Thanks
mental nexus
Thanks all!
still zephyr
Thanks
errant grail
Thanks!
fossil gorge
Thanks everyone. Will catch the replay/recording when it goes up π
thorny jay
Thanks.
gilded cradle
I gotta go feed the cats π±
idle owl
@slender iron I have a quick thing if you have a minute.
We can chat here, it's not personal.
onyx hinge
Here is the notes document for Mondayβs 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 youβll be attending the meeting - itβs super helpful! If you are unable to attend but would still like to include updates, feel free to include them in the notes and weβll read them off during the meeting. Hope to see you there! <@&356864093652516868> https://docs.google.com/document/d/17uYmUOK6HVNQo8EawvMxyTqzSCqq__G5n5Pr8jtGz5A/edit?usp=sharing -- Double check the correct local time, as the US switches to "Daylight Saving Time" on March 14.
slender iron
@onyx hinge ping me if you have time after you do meeting stuff and take a break today re: weblate
tidal kiln
@idle owl hey. got a read the docs / CI question if you've got time.
idle owl
<@!330227457296957440> hey. got a read the docs / CI question if you've got time...
Sure I'm multitasking but I can help.
tidal kiln
it's not urgent. you can ping me later? if ur busy.
fossil gorge
Just a quick note - I'm catching up on the Youtube replay, and audio's quite faint.
turbid radish
?serverinfo
digital shoreBOT
adafruit
onyx hinge
@fossil gorge yeah it's a bit lower than ideal, but I think I at least got me & the audio participant levels closer. I'll continue working on it next time.
OBS had both sides peaking in low yellow which I thought would be okay
they were at -11.2 and -3.3 so I'll try -7.9 and -0.0 for next time
idle owl
@tidal kiln Working on getting the newsletter out, so if that works, I'll ping you later. (I typed this out immediately after you messaged me and got distracted... so yes I'll ping you later.)
onyx hinge
@slender iron I'm back and can chat now
tidal kiln
@idle owl gotcha. no worries.
slender iron
@onyx hinge for the translation "minimum"
I was thinking of using weblate's priority setting
onyx hinge
I'm not familiar with that
onyx hinge
how does one set the priority flag -- in the web ui or in the .pot file?
slender iron
the description I saw was in the web ui
I don't have permissions though
which is why I'm bugging you π
onyx hinge
did you log in with github?
fossil gorge
OBS had both sides peaking in low yellow which I thought would be okay
Sounds good. Not complaining, just wanted to let you know on the off chance. I know OBS and audio in general can be a tad... touchy?
slender iron
ya
onyx hinge
huh, Dan did that and he was awarded admin access right away. I'll get you set up though.
slender iron
maybe I linked it later...
slender iron
yup
onyx hinge
OK you should be added with administration access now
slender iron
thanks!
devout jolt
Hi! Is there a way to get the name of the board (e.g. "QT Py", "Pico", "ItsyBitsy M4") in CircuitPython? It would be handy when creating a single piece of code for multiple devices. I think board.name would be a good place for it. I've checked sys and microcontroller and can only get chip name and CP version
slender iron
I didn't find the spot for it yet
it's under label
onyx hinge
ahhh
slender iron
I'm looking in their github for more clarity
@devout jolt I think there is os.uname() or something too
but ya, it'd be good to provide the board ID too
``priority:N``
Priority of the string. Higher priority strings are presented first to translate.
The default priority is 100, the higher priority string has, the earlier is
offered to translate.
Thanks!. Yes, os.uname().machine will work. Returns 'Adafruit ItsyBitsy M4 Express with samd51g19' for instance.
slender iron
I'll do them as priority:200
onyx hinge
this is cool, we'll be able to use a search like priority >= 200 and untranslated (not sure of exact syntax) to check the status of a language.
slender iron
π
it'll prompt for those first too
k, lunch time. I'll try to mark them all later today
idle owl
@tidal kiln Ok. I have a moment.
tidal kiln
hey! ok...one sec....let me context switch...
tidal kiln
so..as far as i can tell, the RTD here is not current:
https://circuitpython.readthedocs.io/projects/nunchuk/en/latest/
and can't readily see how that gets done in the workflow
onyx hinge
readthedocs build happen on readthedocs, you should be able to see https://readthedocs.org/projects/adafruit-circuitpython-nunchuk/ I think?
idle owl
@tidal kiln Checking.
tidal kiln
right, but what in the CI deploys to there?
idle owl
It looks like it's rendering properly.
onyx hinge
last build was 1 month + 1 week ago
onyx hinge
The github actions CI is not what deploys the readthedocs stuff
github sends a hook or something to readthedocs, and then they build it
tidal kiln
basically, it's not up to date
onyx hinge
https://readthedocs.org/projects/adafruit-circuitpython-nunchuk/builds/12891709/ says it built from ref https://github.com/adafruit/Adafruit_CircuitPython_Nunchuk/commit/19258ca7a6ec8cf148632137ec3acce9168dc60d which looks like it matches the latest per github..
idle owl
You released on 29 January, and I think that's the last time it built.
onyx hinge
is there particular text you're not seeing online?
idle owl
Where is values though?
values property appears to be missing.
Rebuilding manually to see if it's something in the lib that's causing it.
idle owl
It might simply be a glitch. But let me see if it works.
tidal kiln
cook. ok. thanks.
idle owl
Nope.
Let me see.
Hmph.
I'm perplexed.
You're right that it's not updating. values should be in there at least.
tidal kiln
@onyx hinge new values and buttons properties are not showing up. and old ones are still there.
onyx hinge
the metadata in the html say weird things .. "build_date": "2021-03-08T22:18:00Z", ..., "commit": "b653d734",
tidal kiln
@onyx hinge also thanks for looking / helping as well π
tidal kiln
ohhhh. that could totes be it.
tidal kiln
yah. i did that change about that same time frame.
onyx hinge
and somehow, possibly following the link from rtd, I was seeing that "master" when I thought I was on the default branch
onyx hinge
checklist item to add when transitioning a repo to 'main' I guess
tidal kiln
was that a change on RTD or github?
tidal kiln
anything i could/should do? or just bug you when it happens?
idle owl
That did not seem to resolve the issue though....
@tidal kiln Probably bug me? But I should probably sort out not being a single point of failure there. I don't think you do enough with it for it to make sense to give you credentials.
tidal kiln
yep. i agree.
onyx hinge
https://circuitpython.readthedocs.io/projects/nunchuk/en/latest/api.html now has "values"...
idle owl
Oh good then mine is cached.
onyx hinge
"stable" did not appear to change. has that version been released yet?
idle owl
Thanks <insert whatever is caching here>
idle owl
checklist item to add when transitioning a repo to 'main' I guess
Indeed.
tidal kiln
not sure, what designates the "stable" release?
onyx hinge
I'm not sure either.
tidal kiln
"stable" = latest release
"latest" = current repo head
?
onyx hinge
If your versions follow PEP 440, Read the Docs activates and builds the version if itβs greater than the current stable version. The stable version is also automatically updated at the same time. See more in Versioned Documentation.
We also create a stable version, if your project has any tagged releases. stable will be automatically kept up to date to point at your highest version. If you want a custom stable version, create either a tag or branch in your project with that name.
so yes it should be that when a version is given a tag -- that doesn't look like a prerelease per PEP 440 -- it should update the "stable" docs
@idle owl mind prodding the button to re-build the "stable" docs and see if it makes a difference?
idle owl
<@!330227457296957440> mind prodding the button to re-build the "stable" docs an...
Sure.
onyx hinge
or maybe we'll have to tag 1.0.1, to make it really re-think its choices
I wonder if the "release" not being in the history of what it thought was the "default branch" inhibited some of this fancy processing.
idle owl
Rebuilding now.
onyx hinge
https://circuitpython.readthedocs.io/projects/nunchuk/en/latest/api.html has "values" after a shift-reload now
mental nexus
A basic question on .h file organization in shared-bindings and shared-modules. What's the decision point for which function headers go in which location? I assumed that functions in shared-bindings would go in shared-bindings/CPCode.h, but sometimes they're found in shared-modules/CPCode.h.
idle owl
@onyx hinge Thanks for checking since my browser is being a jerk.
onyx hinge
@mental nexus I think that shared-bindings can contain the function prototypes that would be identical in any implementation (thinking of cases where there is a common-hal implementation, for instance). but I am not 100% sure of the rule, I just guess and see what makes it through PR review.
mental nexus
Also regarding function naming, for example fontio. All functions in shared-bindings start with fontio_, but over in shared-modules, some start with common_hal and another one with fontio_ I suspect that the fontio_ ones are exposed to the user while the others are not. But then why is there a fontio_.._get_glyph over in shared-modules. Main reason for my questions is so that I can be clear when adding functions I can do it in the appropriate and expected place.
tidal kiln
yep. looking good for both latest/stable now. thanks @idle owl @onyx hinge
mental nexus
<@!366059101203202052> I think that shared-bindings can contain the function pro...
Ok, understood. I'll look at a few more examples and try and figure out some heuristics to use. Mainly don't want to be swimming upstream...
idle owl
yep. looking good for both latest/stable now. thanks <@!330227457296957440> <@!4...
Great! Jeff did most of the work there. Glad we could help!
tidal kiln
ok. next question. feather rp2040 reset behavior - CIRCUITPY isn't showing up after reset.
idle owl
ok. next question. feather rp2040 reset behavior - CIRCUITPY isn't showing up af...
Did you call show()?
tidal kiln
well. thought about that...but can't....no REPL π¦
this is after copying over latest UF2
all works fine, just nothing after reset
crimson ferry
@tidal kiln not sure if related, but I have seen no REPL, no serial or CIRCUITPY on host (mac)... requires full nuke and re-install to fix (filed an issue)
tidal kiln
@crimson ferry thanks. can you link issue.
crimson ferry
I think itβs flash-related? But could be >1 issue or just something quirky about reset/usb
tidal kiln
@crimson ferry may be related? not sure. at this point I'm not even copying files, other than the firmware UF2 to the RPI-RP2 folder. that gets CIRCUITPY just fine. can connect to REPL. blinked the LED, etc. then simply reset..and...nothing.
hmmm. no CURRENT.UF2 ?
idle owl
@lone axle Do you want to take a stab at updating the datetime library to have another hook like the one for the example code, except for the tests directory?
idle owl
@prime flower I sent you an email regarding the datetime lib. If you have time tomorrow or Wednesday I'd like to chat in Slack or here.
(I would have left it at sending an email but saw your comment in #learn-adafruit-authors )
prime flower
<@!311339734339289090> I sent you an email regarding the `datetime` lib. If you ...
I'm OoO this week (turned on discord to help a friend who doesnt have my # and jerry pinged at the same time), will look then unless its urgent?
idle owl
@prime flower It's not! Sorry to bother you, I forgot you were out. Ping me next week when you're around!
prime flower
np - will do next week! (I use the same discord account for everything, incl school, should really decouple!)
idle owl
@prime flower I totally get it. I'm the same with Discord.
idle owl
@lone axle I commented on the PR with what I'm looking for. If you're not up for it, please let me know in the GitHub thread. Thanks!
crimson ferry
hmmm. no `CURRENT.UF2` ?
something bothered me about that listing, but it didn't register consciously, I have no CURRENT.UF2 either
after bootloader, now reset puts me in never-never land, will need to nuke again
tidal kiln
what is the "nuke" process?
crimson ferry
drop the raspberrypi . org nuke UF2 onto the bootloader, I'll dig up the link
(I haven't found an Adafuit full erase UF2)
link at p.5 here: https://datasheets.raspberrypi.org/pico/raspberry-pi-pico-faq.pdf can't seem to copy the actual link to the file
tidal kiln
what happens if you copy a firmware UF2 without doing the nuke?
crimson ferry
nothing, still won't mount CIRCUITPY, and no indication of serial presence, but code.py seems to be running
that nuker is for Pico, not sure if it's really correct for Feather RP2040, but I'm not aware of one specifically for Feather
tidal kiln
ah, pico doesn't show a CURRENT.UF2 either, so that may not be related or matter
lone axle
<@!382939733107408897> I commented on the PR with what I'm looking for. If you'r...
I think I can do that. I'll give it a try.
idle wharf
@ornate breach might be interested in this as well. I know they were looking into S3 stuff this weekend
ornate breach
It was the C3, unexpectedmaker on the other hand did do a video or a stream with the S3 though π
idle wharf
C3 / S3 lol. π€¦ββοΈ
fossil gorge
C3 / S3 lol. π€¦ββοΈ
They're close enough. I mean, they're nearly ROT13 of each other, so... π€·
lime trellis
@tulip sleet usb_cdc breaks helper scripts that my students use for automatically opening a putty session with the most recent usb device connected (b/c we now have multiple usb ports). What compile-time option do I set to get the old behavior?
tulip sleet
CIRCUITPY_USB_CDC = 0
In general, the lower-numbered port is the REPL port. What is the OS your students are using?
lime trellis
In general, the lower-numbered port is the REPL port. What is the OS your studen...
students manage to dig up every os under the sun π
FWIW I just tested on Windows 10 and my REPL is COM23 with CDC on COM21
I'll comment on the PR with this info so dum-dums like me can find it π
tulip sleet
There are ways to determine which port is which. The interface name is "CircuitPython CDC data" for the REPL and "CircuitPython CDC2 data" for the secondary channel. If we provided a script to identify the port, I assume that would help you. Could you open a new issue, and include the scripts you are using now? My very next task is to get Mu to identify the corretc port.
of course you can use an earlier version of CircuitPython for now unless you need the later version
this is the point of putting it in the beta, to find these cases and figure out what to do π
lime trellis
Thank you for the info!! building now just to verify π
RE older CP: eager beaver students were already building beta 3 and found the hiccup. what can ya do haha
@tulip sleet I've been meaning to ask: are there any efforts to get the REPL running in the browser. Not over wifi, just the browser accessing the serial port?
My biggest CP hurdle with students is a low-effort way of accessing the serial monitor and REPL. Had some push back downloading and installing Mu
tulip sleet
the CIRCUITPY_USB_VENDOR options turns on WebUSB. However, WebSerial would be easier. You'd still need to choose the serial port. I have tried the Google demo WebSerial program and it doesn't quite work, but it's close and probably just needs a little work. In either case I think it may only be supported by Chrome. Gotta go, but feel free to ping me with more q's about this
lime trellis
ty!
onyx hinge
I investigated whether using qstrs as "dictionary words" can improve message compression. It requires moving to a 16-bit representation of chars on all translations, so that the space for "words" is larger Unfortunately, it looks like it's not a net improvement, based on trinket_m0_express de_DE build size.
tulip sleet
thanks for investigating this
manic glacierBOT
manic glacierBOT
manic glacierBOT
manic glacierBOT
fossil gorge
To answer your question from the meeting yesterday Jeff, I'm window-shopping anything! Either a Windows PC (what a pane to find a decent one at a decent price), or a fruitless search for a reasonable second-hand Apple
manic glacierBOT
tulip sleet
@Hugo I have bought many machines from https://www.dellrefurbished.com. There is almost always a coupon of some kind for 30-50% off. See the Coupons & Deals link. The coupons often come out later in the week. Occasionally dealnews.com has private coupons. These are off-lease machines. The "B" condition ones are fine.
I have read that with the advent of the M1 Macs, the price of other used Macs has dropped significantly. But you may not be seeing this.
onyx hinge
huh I get a security warning visiting that url
Websites prove their identity via certificates. Firefox does not trust this site because it uses a certificate that is not valid for dellrefurbished.com. The certificate is only valid for d.sni.fastly.net.
https://www.dellrefurbished.com/ doesn't show a security error
fossil gorge
Thanks @tulip sleet, much appreciated!
For the intel based macs, I think there's some reluctance from folks giving up their current one on the off-chance that the new one doesn't do what they want it, or isn't spec'd "well enough".
tulip sleet
i just got from dell refurbished a 2019 micro-sized Optiplex desktop with 16GB, a 512GB SSD, and an i7-8700T (6/12 cores/threads) CPU, $424 tax included. That was a very well spec'd machine. i5's are much more common. Also nearly all come with Win10 Pro.
fossil gorge
Dang!
tulip sleet
that is after a 40% coupon
onyx hinge
The built-in RTC of this matrixportal I have has a lot of error, a second every few minutes. What is this RTC referenced to and is there any way to improve it? I saw there's a calibration property but neither extreme makes an appreciable difference; the adjustment units of maybe-parts-per-million only have an 8 bit range but the error is on the order of 6000ppm..
tulip sleet
I am not sure the matrixportal has a crystal
fossil gorge
That's pretty much a no-brainer. I just want something that I can put 32 or 64GB of RAM in, so I can Docker and VM things I need, but wow...
tulip sleet
it was a 16GB SODIMM, so there's room for another. I would have preferred a SFF size, which has more USB ports, but I really wanted the 6 cores
onyx hinge
this is while connected on USB, I know some devices have OK timekeeping on USB and really poor timekeeping off of it.
tulip sleet
you can also find stuff like this on eBay, but the prices vary more
onyx hinge
and probably the dependability of the product too
tulip sleet
i have had some laptops fail but no Optilex desktops
fossil gorge
I'm reluctant to go the eBay route, but "official" or well refurbished products I'm quite ok with
lone axle
@idle owl when you have a moment sometime can you look into the DisplayIO_Layout library docs. This is the latest merged PR on it: https://github.com/adafruit/Adafruit_CircuitPython_DisplayIO_Layout/pull/9 As far as I can tell all of the actions seemed to pass successfully. When I build and look at the docs locally they look good to me, and do include all of the new things. But the page that is live on readthedocs when we follow the badge link seems to be out of sync with the repo. It's showing an older copy of the docs before some of the more recent additions were made. DanH took a look last night and found a webhook URL that needed updating, but we are still seeing that older copy on the live page. (Perhaps it would need another PR to be merged now that the setting was changed though for the live page to actually update?)
tulip sleet
before that my dev machine for three years was an Optiplex 7010 i5 with SSD I got for $100 on craigslist
fossil gorge
Wow!
tulip sleet
i added some RAM and another SSD but it's still fine. The new machine is about twice as fast on multicore CircuitPython builds, but the raw speed of a single thread is not that much greater
I use the "Compare" option on https://cpubenchmark.net a lot
manic glacierBOT
idle owl
@lone axle For some reason "latest" didn't rebuild when you did the release. "stable" shows the latest things. I did a manual rebuild on latest, but it shouldn't be required.
idle owl
<@!330227457296957440> when you have a moment sometime can you look into the Dis...
It shows up properly now for latest.
lone axle
It shows up properly now for latest.
Awesome! thank you!
idle owl
@lone axle Keep an eye on it though. Because it should have updated.
lone axle
Will do.
idle owl
@lone axle or @fossil gorge Do one of you want to handle this? I'll tag/assign you if so. https://github.com/adafruit/cookiecutter-adafruit-circuitpython/issues/113
fossil gorge
I can handle it, since @lone axle has a few other things going on
idle owl
Great, thank you!
fossil gorge
Any time! π
lone axle
I can handle it, since <@!382939733107408897> has a few other things going on
Thank you!
idle owl
What's another word for "capability" in the context of GPIO pin capabilities
I don't like that word in what I'm writing up, but I'm completely blocking on another option
onyx hinge
I guess I need to join the ranks of people having trouble with mass storage on the feather rp2040. I've been trying to copy some large-ish "mp3" and either it's hanging or it's taking much longer than I expect.
blissful pollen
abilities?
onyx hinge
the repl seems to get confused too
fossil gorge
What's the sentence or context?
tulip sleet
are you using beta.3 or latest? Try a newest latest because Scott fixed a tinyusb regression
idle owl
It's a section title in the Feather RP2040 guide, and it's going to be pins listed by what they can do, versus by pin name. So the title is "GPIO Pins by Pin (Capability)"
tulip sleet
oops, it's not merged yet ,let me merge
solar whale
@idle owl -- would "function" or "functionality" work?
fossil gorge
Functionality?
onyx hinge
looks like I have a random old ref from my rotaryio work, I'll update
idle owl
@fossil gorge @solar whale Thank you.
blissful pollen
was goign to say features, but functionality is better
fossil gorge
The human thesauruses... thesauri? thesauriae?
blissful pollen
the pin's "super powers" would also be an okay choice π
fossil gorge
or "Spells available"
idle owl
Hah! If the pin concepts for this weren't already a bit confusing π
fossil gorge
The pin number, that maps to the GPIO number available with the spells and/or superpowers of blinking an LED...
mental nexus
<@!382939733107408897> Keep an eye on it though. Because it should have updated.
Thanks kattni and foamyguy!
blissful pollen
Hah! If the pin concepts for this weren't already a bit confusing π
Maybe too advanced for the guide but one thing I ran into for the ParallelBus display was using consecutive pin numbers for some devices, have to be based on the actual RP2040 GPIO #s
fossil gorge
Just a thought, might it be worth having both pin numbers and GPIO numbers in the board for the PiPico? so you could have board.PIN25 and board.GPIO13 (I don't have the datasheet handy, so made up an equivalency). And then the GPIO is defined as, essentially, board.GPIO13 = board.PIN25 (or vice versa).
Because there is a lot of crossing of terms in many resources I've seen online so far, and it's always prefaced with "Make sure you're using the right PIN/GPIO number"
Not sure what impact it would have on space thought.
onyx hinge
>>> import storage
>>> storage.erase_filesystem()
``` OK .. umm .. storage.erase_filesystem() still doesn't work !? It acts like it does (clean reboots) but filesystem contents aren't re-initialized?>>> import os; os.listdir('.')
['.fseventsd', '.metadata_never_index', '.Trashes', 'code.py', 'lib', 'boot_out.txt', 'hello.txt', 'alot-128.mp3']
>>> import storage; storage.erase_filesystem()
[tio 11:07:57] Disconnected
[tio 11:08:03] Connected
Auto-reload is on. Simply save files over USB to run them or enter REPL to disable.
Press any key to enter the REPL. Use CTRL-D to reload.
Adafruit CircuitPython 6.2.0-beta.3-75-g7d92f5aa5 on 2021-03-09; Adafruit Feather RP2040 with rp2040
>>> import os; os.listdir('.')
['.fseventsd', '.metadata_never_index', '.Trashes', 'code.py', 'lib', 'boot_out.txt', 'hello.txt', 'alot-128.mp3']
I thought the same thing when I was having to reference GPIO to make sure I had the 8 consecutive pins in a row. Not a big deal. But since the GPIO is on the bottom silk may be useful to have
onyx hinge
ooookay, that's my fault, but why? well, we can revert it.
idle owl
The GP pins are listed so folx who want to use it with the SDK can. We don't intend to add them to the board def for CircuitPython, as far as I know.
onyx hinge
still having trouble copying files π
idle owl
@tulip sleet Opinion time. The PWM pins for Feather RP, includes PWM0 A: PWM0 B: etc, up to PWM6 (because PWM7 isn't exposed). Every one of them except one have pins exposed. PWM2 B does not. Do I leave it in the list with "(none)" after it or remove it from the list since there's no pins for it?
tulip sleet
i think none is good because then it's obvious it's not a typo in the list
idle owl
Ok thanks
@tulip sleet Was that what you had in mind?
When you suggested pins listed by function in your review of this page last week
onyx hinge
from the fab print I can't figure out the debug header orientation.. https://learn.adafruit.com/assets/100338
OK from the original schematic I can see it, there's a pin 1 arrow on the upper left corner, not visible in the fab print for some reason
manic glacierBOT
slender iron
still having trouble copying files π
did you make sure you got the updated tinyusb again?
onyx hinge
yeah, I did.
slender iron
Just a thought, might it be worth having both pin numbers _and_ GPIO numbers in ...
Note, the GPIO numbers are always available in microcontroller.pin
onyx hinge
if I hadn't, the version number would have said "-dirty", I am pretty sure
slender iron
tinyusb should be 2ad
tidal kiln
should i open a new issue for this?
https://github.com/adafruit/circuitpython/issues/4349#issuecomment-794003709
fossil gorge
Note, the GPIO numbers are always available in `microcontroller.pin`
Good to know, will try to remember that. I keep getting proverbially turned around with the Pico for some reason
slender iron
@tulip sleet beta.3 had the mistaken tinyusb rollback in it right?
manic glacierBOT
tulip sleet
<@!329766224093249548> beta.3 had the mistaken tinyusb rollback in it right?
No, it was in the .transpose_xy fix, which is in pr #4331, which was after beta.3.
tinyusb commit in 6.2.0 beta.3 is https://github.com/hathach/tinyusb/tree/280297bdb7aec67adf347ec046943a48a71647df
tulip sleet
@slender iron so there were two rollbacks of tinyusb??
slender iron
no, just one
tulip sleet
since 280927, there have been some interesting rp2040 fixes in tinyusb, like https://github.com/hathach/tinyusb/pull/700
slender iron
right, that's why I was thinking beta.3 was bad
but those fixes were merged after beta.3 and then rolled back accidentally
tulip sleet
no, maybe not that, but any way, a bunch of new commits
tulip sleet
yah, and maybe we should update tinyusb to latest
manic glacierBOT
tulip sleet
we just merged https://github.com/adafruit/circuitpython/pull/4371, which is even newer, and reverts another change, not tinyusb-releated
slender iron
right right
manic glacierBOT
So, I just used GitHub on the web to update the PIDs for the bootloader and this at the same time. (Literally was going through the edit/commit/PR flow on both sides. Somehow, Github managed to commit this file despite being in the other repo. Look at this:
Not that it matters at all, but really crazy. Never seen that!
lavish saffron
HI, how does my community bundle driver get uploaded to PyPi? The Build CI runs twine check dist/* but I can't see a step that actually uploads it. Thx.
manic glacierBOT
tidal kiln
@lavish saffron the pypi deploy happens from the individual library repo, not the bundle repo
lavish saffron
<@!369681896034009101> the pypi deploy happens from the individual library repo,...
Hi, thanks for getting back to me. I'm sorry, I was looking at the on push instead of the on release actions. I can see an authentication error now and I'll try to fix it π€¦ββοΈ
lavish saffron
Hi, the upload-pypi release action is failing with Invalid or non-existent authentication information. See https://pypi.org/help/#invalid-auth for more information. I have checked my PyPi API token in the GitHub repo's Repository secrets as per the link in the error message and it seems fine, so I'm not sure where to go from here?
manic glacierBOT
still zephyr
@lone axle Very good job in the New display_text guide. I love it!!
lone axle
Thank you!
still zephyr
@lone axle question here I saw that we have the Refactor in, do we introduce changes rebasing on this and canceling olds PRs?
lone axle
@onyx hinge I think there may be a permission issue with the notes doc for next week linked in the pinned message. When I follow that link it takes me to a page to 'Request Access'
still zephyr
yes same here for the doc for next week
lone axle
<@!382939733107408897> question here I saw that we have the Refactor in, do we i...
I don't have much experience with rebasing. I would probably lean toward merging adafruit/master into the branches from the PRs and resolving any conflicts if there are any. I am totally willing to help do some of that work as well if you want.
Sorry that the refactor came right in the middle of those PRs and require more tweaking on them.
Or starting a new branch if you want for clean commit history, instead of merging. But it's no big deal to have those commits in either I think.
still zephyr
mmm, do not worry, not saying that I do not want help :), but it is ok, I have all the changes in my mind. And I prefer to start a new branch as you mentioned. I am open, if you want to help, just let me know which one do you want to take, I do not want that we work on the same PR π
But after yesterday C rabbit hole a little of python fresh air is most welcome for me
lone axle
mmm, do not worry, not saying that I do not want help :), but it is ok, I have a...
New branches sounds good. So yep we can close those PRs and make new ones when the new branches are ready. I'll do the font example one if you want.
still zephyr
Ok deal. Thanks @lone axle I will work on the others. Thanks again
onyx hinge
@lone axle I'll fix it when I'm back at a desktop computer.
I got it wrong at first and trying to fix it made it worse
marble hornet
@mental nexus did your slide switch ever make it into a library on gh? would love to try it out and having some trouble finding it
onyx hinge
<@&356864093652516868> the permissions on the notes document for next week have been corrected -- please let me know if you still have trouble @lone axle @thorny jay
lone axle
<@!366059101203202052> did your slide switch ever make it into a library on gh?...
It was merged in yesterday. It's available now: https://github.com/adafruit/Adafruit_CircuitPython_DisplayIO_Layout
lone axle
<@&356864093652516868> the permissions on the notes document for next week have ...
all good now. Thank you
marble hornet
It was merged in yesterday. It's available now: https://github.com/adafruit/Adaf...
thanks! I'll give it a try
mental nexus
Yep, and I tried to be diligent on the docs for this first one so I wouldn't forget what I did whenever I came back to it. Let me know if you find things for improvement. https://circuitpython.readthedocs.io/projects/displayio-layout/en/latest/api.html#switch-round
onyx hinge
idle wharf
@analog bridge are you around ?
analog bridge
<@!747673663419187281> are you around ?
yep!
idle wharf
Question for you. Does your update-idf branch build successfully?
analog bridge
idle wharf
Mine does not. Why don't you just do a PR and I'll close mine.
analog bridge
Mine does not. Why don't you just do a PR and I'll close mine.
I think I can commit directly to your branch... have you enabled the Allow edits by maintainers tick box on your PR?
idle wharf
Yes that is checked.
ONe thing to note. I tried pulling in your changes AND adafruit main and both times, I ended up with the same error outside the eps32s2 port
QSTR updated
../../supervisor/shared/workflow.c: In function 'supervisor_workflow_connecting':
../../supervisor/shared/workflow.c:39:12: error: implicit declaration of function 'tud_connected'; did you mean 'tud_connect'? [-Werror=implicit-function-declaration]
return tud_connected() && !tud_suspended();
^~~~~~~~~~~~~
tud_connect
../../supervisor/shared/workflow.c:39:12: error: nested extern declaration of 'tud_connected' [-Werror=nested-externs]
cc1: all warnings being treated as errors
this must be related to submodule update
idle wharf
Possibly... I just wanted to mention it.
Also if you don't want the $(IDF_TARGET) mixed in, submitting your own PR would clean that up as well.
But I see you didn't update the IDF's commit, you pointed the submodule to espressif's repo\branch instead.
analog bridge
Also if you don't want the $(IDF_TARGET) mixed in, submitting your own PR would ...
yes I am following the release/v4.3 branch
have you got a backup of IDF_TARGET changes as I will be rebasing your branch
idle wharf
They are super easy to do again now that I know how, go for it.
analog bridge
okay... something came up just now... I'll get back to this shortly
idle wharf
No worries, Thank you MicroDev !
analog bridge
π
idle wharf
Thanks again @Micrdev We should have just done this 20 days ago π
Also, you were spot-on, once I pulled your changes, I had to update submodules again and now I'm building successfully.
analog bridge
nice... I'll rebase again... my local pre-commit checks were turned off which led to the ci failure.
analog bridge
@idle wharf I might have overwritten a commit of yours...
analog bridge
oh...π
idle wharf
Boom... esp32s2 port on IDF 4.3 builds clean thanks to @analog bridge !
analog bridge
@idle wharf I am taking a look at $(IDF_TARGET) now...
have you got i2c working with v4.3?
idle wharf
Yes. I was able to run the 4.3 based build on a feathers2 and access a bme680 over i2c.
analog bridge
that's interesting cause bruce mentioned some i2c semaphore issues
idle wharf
Earlier today, I saw issues with I2C in the build, but once I merged in your changes in i2c.c\i2c.h it got passed that point.
simple i2c environment sensor program
ben running fine for about 40-50 minutes
analog bridge
ok... I'll test it, SPI does crash though... I am running a backtrace now
idle wharf
I can scan I2c fine too ...
analog bridge
I think I got why SPI crashes... fixing now.
idle wharf
I don't think it was a question, but UART is working... (I had a GPS lying on my desk so I hooked it up.)
onyx hinge
I wonder which other optimization flags are in use on that particular file in the micropython tests .. like, we've already gone to -Os there for size savings in the builds where we enabled these new size optimizations
but reviewing the discussion around the 5x speedup/slowdown, that's when background tasks were heavyweight too
so maybe we should re-test that
tulip sleet
@onyx hinge I did look at the build for that board, and it is -Os mostly
manic glacierBOT
solar whale
For anyone interested -- just hooked up a rfm9x radio breakout to my RP2040 Pico via SPI -- works normally. No issues transmitting or receiving.
tulip sleet
the background stuff netted another 2x or so
manic glacierBOT
idle wharf
@supple gale if you have some time, checkout the udate-idf PR it would be good to know if you have any I2C issues.
I did not, but you mentioned you did on Microdev's fork previously.
slender iron
@uncut nexus how are you doing wrt pulsein? it still needs another iteration at least. I can pick it up if you are getting tired of it
timber mango
For anyone interested -- just hooked up a rfm9x radio breakout to my RP2040 Pico...
Hi Jerry cool just for testing issues or do you plan to replace the s2, I took a look on the pico, but for me, on a lipo perspective, i will stay with S2
solar whale
@timber mango I was just experimenting - I like to verify that I can use the rfm9x boards with all microcontrollers so I keep my options open. There should soon be some low power options for the nRF52840 boards so I'll also want to try them again once it is released.
manic glacierBOT
timber mango
Hi makes sense, you mean the board of nordic? A good thing for would be the adafruit rfm95x support pi.alarm with deep sleep. On Ffriday my case for the s2 will hopely arrive π so I can attach it to the door π
manic glacierBOT
manic glacierBOT
supple gale
<@!342132923064909840> if you have some time, checkout the udate-idf PR it would...
Was on my list as the next thing to do.
simple pulsar
Hi, I've got some minor displayio weirdness on a no-brand SSD1306 attached to a Pi Pico via i2c. It was working well until I introduced a Bitmap and some single pixel updates. It now pauses sometimes and corrupts a few pixels in an unrelated Rect. What's most likely part of all of that to be buggy? I dropped i2c to 100kHz and that made no difference.
tough flax
Got my Feather RP2040 today & tried to make an AT HeadMouse using a BNO055 - it worked... briefly π
I made a video: https://www.youtube.com/watch?v=6dau3qSTxW8
Are there known issues w/I2C on this board?
I will go look at issues, but was hoping someone could point me in the right direction
lavish saffron
Hi, the GitHub upload-pypi release action is failing with Invalid or non-existent authentication information. See https://pypi.org/help/#invalid-auth for more information. I have checked my PyPi API token in the GitHub repo's Repository secrets as per the link in the error message and it seems fine, so I'm not sure where to go from here?
idle wharf
Hi, the GitHub upload-pypi release action is failing with Invalid or non-existen...
Is your repo public?
Never mind, found it π
https://github.com/BiffoBear/CircuitPython_AS3935/actions/runs/634102083
I think something in your setup isn't quite right...
If you look at this https://github.com/BiffoBear/CircuitPython_AS3935/runs/2071312909?check_suite_focus=true#step:6:7
You'll see the user name and password are blank.
Run python setup.py sdist
python setup.py sdist
twine upload dist/*
shell: /usr/bin/bash -e {0}
env:
pythonLocation: /opt/hostedtoolcache/Python/3.9.2/x64
TWINE_USERNAME:
TWINE_PASSWORD:
But if you look at this successful run from another library
Run python setup.py sdist
python setup.py sdist
twine upload dist/*
shell: /usr/bin/bash -e {0}
env:
pythonLocation: /opt/hostedtoolcache/Python/3.9.2/x64
TWINE_USERNAME: ***
TWINE_PASSWORD: ***
There are placeholder "***" there which implies in your case the values are not being picked up.
silver tapir
Ok, weblate (-es) back to 100%
manic glacierBOT
onyx hinge
Is it OK to merge a PR while CI is failing like this, or should I wait until the test/ lint changes go in? https://github.com/adafruit/Adafruit_CircuitPython_datetime/pull/6 -- I think the change we want to make hasn't actually been made in cookiecutter or added as an adabot patch, so if I do something now I risk making it harder to apply the patches later..
@silver tapir thanks!
lone sandalBOT
lavish saffron
I think something in your setup isn't quite right...
If you look at this https:...
Thank you for your reply. I have my PYPI API key saved as a secret, but based on your information I can see that the release.yml file looks in an environment for ${{ secrets.pypi_username }} and ${{ secrets.pypi_password }} . I've created an environment and added the secrets with the username as __token__ and the password as my API key. I still get the same error though. Is there a naming convention for the env, or permissions that I have to set to allow the release.yml to access the secrets?
lone axle
Thank you for your reply. I have my PYPI API key saved as a secret, but based on...
I was able to get the pypi action working by creating these two "secrets" in the settings page of my repo:
they are specifically under "Repository Secrets" not "Environment Secrets"
tulip sleet
@slender iron re our ChromeOS discussion yesterday, do you have more to add to this https://forums.adafruit.com/viewtopic.php?f=48&t=171766, like the Google bug and resolution pointer?
gloomy shuttle
<@252717193496756235> re our ChromeOS discussion yesterday, do you have more to ...
FYI, I had issues last fall with Chrome browser (not just chromebooks) and serial connections with Arduinos and https://duino.app. Several chrome updates kept breaking it, but the developer was able to resolve the issue.
ionic elk
@onyx hinge do we want to close that issue? canio should work on every F4 board that has it, I don't see a reason to keep it turned off unless you had problems in practice on other boards
The F407 in particular is essentially identical to the F405
manic glacierBOT
manic glacierBOT
slender iron
@tulip sleet the issue I've been watching is serial only
it feels more like a google issue than an adafruit one
tulip sleet
i agree, they are not liking the USB drive for some unknown reason
the weird thing is it seemed to be fixed, and then regressed. I will have to break out the Chromebook again and try it before and after an update
slender iron
https://bugs.chromium.org/p/chromium/issues/detail?id=1169000#c3 is the serial issue
ya, I suspect its just something they don't test well
tulip sleet
do you have any kind of inside track?
tulip sleet
well, I will see if I need to open another issue :/
slender iron
ya, that's probably best
I think the usb allowlisting may only impact thonny as well
because it uses pass through or something
slender iron
anyone done webserver stuff on the ESP32-S2?
onyx hinge
@ionic elk I don't have any of the HW, so I didn't even try. Each board needs lines added to ports/stm/peripherals/[...]/periph.c at a minimum. And that assumes the C API is the same for all the various chips. I am fine leaving it open if you want to interpret it as a feature request, but as written it was a support request and that's been answered.
ionic elk
@onyx hinge it's no biggie, the F407 is a strict superset of the F405 so I put in a PR that updated it, since it was lagging behind the F405. The F412 would have required actual work so I ignored that part.
onyx hinge
@ionic elk thanks! I would not have known any of that. I just have vague warnings that they can be different in unexpected ways
ionic elk
Yeah the F412 in particular had weird register level changes for no reason
if you only use the HAL things tend to be consistent within a family but the H7 and F7 HALs change stuff in really subtle breaking ways so those take a lot of double checking.
The F407 is a specific case where it's actually literally the same silicon as the F405 with more features, they even share the same datasheet.
tulip sleet
@mental nexus I have an uninformed question which is whether bitmaptools and vectorio share any purpose. They both can define rectangular regions, for instance.
mental nexus
As I started exploring the capabilities of vectorio just this week, I am now becoming less than clear on which is best to use. Here's what I know so far. For rotating, scaling and placing text, bitmaptools is currently the only option available.
As for manipulating polygons, we have a couple of libraries available, the built-in vectorio and the display_shapes library.
tulip sleet
when I was casually reviewing the latest additions to bitmaptools, it seemed like there was an overlap, and I was wondering if a unification was possible. vectorio does not deal with Bitmaps. Anyway, it's something to think about in the long run, maybe
mental nexus
Yep, I agree and the boundaries are overlapping (no pun intented). I think if you want to build a stagnant background, then bitmaptools may be the best way to go. If you have shapes you want to move their relative positions around, then vectorio or display_shapes is probably a better approach.
tulip sleet
got it, bitmaptools does not deal with "objects". It paints
mental nexus
Also, display_shapes is built in python, so it's relatively slow. You could probably build the same thing with bitmaptools if anyone wants a speedup.
tulip sleet
thanks - i have basically not used displayio and friends in any informed way
and so was wondering
mental nexus
One good thing about bitmaptools is you can pre-allocate all your bitmaps and keep them around and paint in them. That way you can avoid the memory segmentation issues if you keep your bitmap buffers around.
I was spurred into thinking about the differences in vectorio and bitmaptools and display_shapes by todbot and his dial gauge this weekend. I think vectorio is an underappreciated library. It's relatively fast at redrawing and pretty flexible.
low sentinel
@mental nexus OOM making layered static backgrounds with bitmaps is what drove me to make vectorio. Its api is pretty ready to add things like clipping masks and rotation in C. Just havenβt yet had a need.
Bitmaps take a huge amount of memory compared to a microcontroller when making an interactive UI. I reach for them when I actually need bitmaps. When I need a colored square or circle, itβs nice to have constant memory space requirement.
fossil gorge
I was spurred into thinking about the differences in `vectorio` and `bitmaptools...
There's an issue on the ProgressBar library, maybe even opened by you, to move to the vectorio library instead of bitmap, for memory savings.
mental nexus
<@366059101203202052> OOM making layered static backgrounds with bitmaps is what...
WoW those are great points. Just to note for folks that with your vectorio you can already do polygon rotations by redefining a polygon's points (and I guess you could write your own wrappers to input angles rather than points). In @devout jolt 's dial, one key performance limiter is redrawing a fairly large rectangle when the dial needle goes to a 45 degree angle. In vectorio is there any way to make an arbitrary "dirty polygon" for speeding up redraws of angled shapes like a dial needle?
Oh, and vectorio is pretty good about filled shapes too. I think I'll gravitate toward using it more in the future.
low sentinel
Itβs easy to write rotations in python. You can pin your coordinate system (in the vest or shape) at 0,0 and do the point math to rotate your point array about that axis.
It tries to avoid doing work it does not need to, what dirty area are you talking about though?
mental nexus
When drawing a rectangular dial needle, when it's horizontal, the dirty rectangle is the same dimensions as the needle. As it moves up to 45 degree angle, the "dirty" boundary redraw rectangle to cover it gets really big (meaning it has a lot of pixels more than just the skinny needle rectangle). So the display is asked to update a lot of pixels that aren't really "dirty".
My understanding of how displayio works is that it asks for every dirty pixel to be redrawn.
low sentinel
Thereβs logic in vectorio polygons so you could place your shape at -10000,-10000 and place polygon vertices within it at 10100,10050, and it will only mark the potential size of actually possible pixels not some gigapixel sized area.
Yeah Iβm familiar with that however thereβs not really a general way to do area marking like that without marking off the area.
Dirty rectangles life be like
Could use a gpu to open doors but weβre running on uCβs. I do quite like EVE2 but itβs quite a leap from displayio to that
mental nexus
Thanks for the insights on vectorio. As I start exploring it more, if I run into things I don't understand, I may pepper you with more questions. It's a great set of features, so thanks for adding it in there. Now, I'll just have to work understand the best ways of using all the capabilities it has.
manic glacierBOT
Linux:
and Windows 7.
Same behavior for me on both.
I have noticed that code.py does seem to persist. I saved basic blinky to it, and that same script was still there and ran just fine after subsequent re-pasting of the UF2.
meager fog
@lone axle hihi
lone axle
<@!382939733107408897> hihi
Hello π
meager fog
ok for the jp project, we're going to use the 3.5" TFT featherwing
dya have an rp2040 feather?
lone axle
I do, my two came yesterday.
meager fog
ok so ya just need the 3.5" TFT wing?
lone axle
Yep, that is correct. I did put in an RMA order for that a little bit ago.
lone axle
I did order it, maybe an hour or two ago.
meager fog
ok what was the order number
lone axle
Order Confirmation No: 2509494-2006762344
meager fog
ok thanks
alright so the project part ya can help with is making nice lil image-buttons
with nice labels π
i figure you're doin a lot of it now
lone axle
Yep π JP shared the main ideas with me. I think some the new stuff in the DispayIO_Layout library will be perfect for making little touchable icons with labels in a grid.
meager fog
ok thankx!
lone axle
For sure. Happy to help out with it! Thank you.
thorny jay
@meager fog did you get my suggestion to make your NeoKey pluggable at the back of a QT-Py (like todbot example with rotary encoder)?
thorny jay
Just if the pin space is the same and the pin match... you have a one key keyboard.
like this https://twitter.com/todbot/status/1365381967013797888?s=20 (It felt like the "NeoKey" is about the same size as the QT-Py.
onyx hinge
uint8_t fontio_builtinfont_get_glyph_index(const fontio_builtinfont_t *self, mp_uint_t codepoint) huh I guess nobody noticed that the builtin font could only have 256 (255?) distinct glyphs
manic glacierBOT
manic glacierBOT
lavish saffron
Thank you for your reply. I have my PYPI API key saved as a secret, but based on...
Well, I'm finally there. Thanks for pointing me in the right direction π
gaunt spruce
Hi all qq, Im working on a MatrixPortal M4 board using circuit python and I have my secrets file working able to access the internet, but for whatever reason the adafruit_requests library isnt working. Specifically Im, importing adafruit_requests as requests and trying to perform r = requests.get('https://api.coindesk.com/v1/bpi/currentprice.json') I keep getting the following traceback error:
Traceback (most recent call last):
File "code.py", line 12, in <module>
File "adafruit_requests.py", line 681, in get
AttributeError: 'NoneType' object has no attribute 'request'
Any thoughts, normal requests package this works just fine
manic glacierBOT
idle wharf
Hi all qq, Im working on a MatrixPortal M4 board using circuit python and I have...
I'll reply on #help-with-circuitpython
still zephyr
@vague latch Regarding the Jupyter Notebook, actually Python in the computer are looking for the information in the ports and then, acting according the OS to look for the possible USB connected devices. In the CircuitPython side as I mentioned before, normally we could use the uname function in the OS module. So we do not have a list per se, the list is generated on the fly, when we request the serial communication from the jupyter notebook to the circuitpython board, but I agree with you, we could make some decision either in the Jupyter notebook or on the board
I did some work in the past in micropython and the jupyter kernel (only as a User), but I am wiling to explore the CP possibilities know
still zephyr
in circuit python on the boards
vague latch
so if the board isnt supported it wont have uname
so it would be better running this inside jupyter
still zephyr
is not supported by CircuitPython?
Yest it would bebetter, but we have the list of boards avalaible that we support
vague latch
we use the look up to check if the device is supported right
cedar moth
Good afternoon all. I just received my Feather RP2040 and am trying to get it to work with an SH1107 128x64 OLED Featherwing via I2C. I am getting an error when trying to initialize the I2C connection using the following code: `Adafruit CircuitPython 6.2.0-beta.3 on 2021-03-04; Adafruit Feather RP2040 with rp2040
import board
import displayio
displayio.release_displays()
i2c = board.I2C()
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
TimeoutError: Clock stretch too long
`
vague latch
yay i2c
cedar moth
Any thoughts appreciated, or should I file an issue?
crimson ferry
still zephyr
I am thinking that you need to use busio
vague latch
@still zephyr as board.py looks at the serial port. I dont think we would want to muddy this with os.uname(). As that is circuit python specific.
cedar moth
Thanks @crimson ferry and @still zephyr ... that was quick! I will give it a try. That code works on a Feather M4 Express and a QT Py so I thought it would with the RP2040.
still zephyr
@vague latch yes agree, we would need to change board, in order to this to work, maybe using some of the attributes of the port object
Maybe rethink _find_board(), how many devices are you dealing with?
vague latch
For me i only have a few. But i want to have the system work for anyone out of the box
There isnt much we can do with the port object. https://pyserial.readthedocs.io/en/latest/tools.html#serial.tools.list_ports.ListPortInfo
still zephyr
π§
But in order to the CircuitPython Kernel to work it needs to be connected to the REPL, so that could be our Try... Except
just tell the user that is not yet implemented
vague latch
@still zephyr thats probably the simplest. We dont need to check the board at all. We just need the confirmation that the board is connected.
still zephyr
@vague latch do you have your project in Github?
vague latch
not at the moment. I am also working on adding a %magics to the kernel to allow python code to run as normal.
Not sure how to delineate them at the moment. Or whether the user should do that them selves
still zephyr
in CircuitPython kernel?
vague latch
yeah
vague latch
My idea as was mentioned by someone else. is to utilise the amazing software library nbdev. This allows you to run the repl for testing and then run nbdev functionality to port your code to a .py file
You can also add documentation inline
still zephyr
Let me know if you need a hand. I would like to do more stuff with CircuitPython in the Jupyter Notebboks
vague latch
nbdev finds all code that has the %export magic at the top of the cell and ports these to a file.