Also, at some point I was hoping to chat with you briefly about timekeeping, basically monotonic time vs RTC time
#circuitpython-dev
ionic elk
tulip sleet
i don't know it that well, maybe in the 1pm or in the weeds?
ionic elk
Sure it might be a good in the weeds topic
1pm I don't like to do anything other than lurk since it's so time constrained
tulip sleet
re tabs, I would think there would be an uncrustify option to untabify
we can ask jun2sak not to use tabs also
ionic elk
ok I haven't read the uncrustify docs, just the yaml file
tulip sleet
me neither
ionic elk
@tulip sleet sure. Do we have a style doc for contributors anywhere?
tulip sleet
no, and I think the idea is that uncrustify will just fix it up anyway
though things like naming styles it obviously will not
ionic elk
right, just automate style, and don't sweat what can't be automated
there was something else I wanted to add to uncrustify... what was it...
maybe comment blocks and long function params? I'll read the manual.
tulip sleet
we are trying to use the same format as micropython so the merges will be easy
idle owl
Easi_er_ 😉
jaunty juniper
I2C.writeto says it expects a ReadableBuffer and takes strings, is a string a ReadableBuffer ?
manic glacierBOT
idle owl
@still zephyr Are you still available sometime tomorrow to chat about your work with aggregating the Adafruit and Community libraries?
tulip sleet
`I2C.writeto` says it expects a `ReadableBuffer` and takes strings, is a string ...
I am looking for the def of ReadableBuffer and having a surprisingly hard time. It's definitely defined _somewhere).
tulip sleet
`I2C.writeto` says it expects a `ReadableBuffer` and takes strings, is a string ...
It should not actually take strs. But https://github.com/python/typeshed/blob/4d734e38dde295e93c27efd6ead3bb5c05d03fe5/stdlib/_typeshed/__init__.pyi#L169
tulip sleet
it is just for typechecking annotation. It's not just a convention, it can be checked
in MicroPython, string and bytes share a lot of code
and sometimes it's too easy to substitute one for another
jaunty juniper
I was looking at the OSError: [Errno 19] Unsupported operation error with the SerLCD and at first I seemed to have different behaviours if I called writeto with a string or a bytearray, but that was a fluke, but still I was wondering about that
tulip sleet
ideally it should validate the arguments as obeying buffer protocol but not a string
onyx hinge
<@&356864093652516868> We're about 1 hour out from our weekly meeting -- please add your notes to the document! See you in the voice channel shortly. https://docs.google.com/document/d/1IBhFVC1LdW_ptX07F6KpWOSZBpEv2eZWFguEIP4nJio/edit?usp=sharing
still zephyr
<@!607375730610798634> Are you still available sometime tomorrow to chat about y...
Hello yes I am 🙂 I took a look during the weekend to better understand,so yes no problem for tomorrow 🙂
idle owl
Hello yes I am 🙂 I took a look during the weekend to better understand,so yes n...
Excellent! Wanted to verify before adding it to my plans.
still zephyr
No problem, let me know around what time and it is all good
idle owl
No problem, let me know around what time and it is all good
Text chat or did you want to do audio/video?
Either is fine, simply trying to plan for it.
still zephyr
I could type pretty fast, but for this I prefer audio
idle owl
I could type pretty fast, but for this I prefer audio
Alright, perfect. I'll figure out a couple of options for you to choose from and we'll figure it out sometime today.
still zephyr
Alright, perfect. I'll figure out a couple of options for you to choose from and...
Thanks. let me know, just to run accordingly, I have the day off so no problems there
idle owl
Excellent! Thank you!
solar whale
@onyx hinge I have to miss the meeting this week. Sorry no time to add notes.
idle owl
Thanks. let me know, just to `run` accordingly, I have the day off so no problem...
How about tomorrow 15:00 US EDT (UTC -4h)?
still zephyr
How about tomorrow 15:00 US EDT (UTC -4h)?
Perfect.. works for me
lone axle
I think the linkified "this post" in the new version of Mu news item in the meeting doc might be going to the wrong place. Currently it leads to a blank new issue page on github.
idle owl
Perfect.. works for me
Great! Added to my calendar.
slender iron
brb
idle owl
same
mental nexus
Special request: can someone change the pinned link for memory saving in the help-with-CircuitPython channel to the new learn guide? https://learn.adafruit.com/Memory-saving-tips-for-CircuitPython. I think it’s a better starting point than the GitHub page.
onyx hinge
@mental nexus I think that you can edit your message and the pinned message text will update too
try that first, and let us know if not
mental nexus
Cool. That worked. Easier than I thought.
idle owl
So with the audio issues I've had, it turns out Discord doesn't support the hardware setup I have at all. So the fact that it even works as well as it has is a total fluke.
thorny jay
Is CedarGrove in this place? What would be his nickname here?
errant grail
Hello @thorny jay !
thorny jay
Hello <@!681528999520108564> !
We can talk thermal camera later. 🙂
errant grail
We can talk thermal camera later. 🙂
Absolutely!
fossil gorge
About this
GitHub made it so first time contributors require someone to “approve” the CI to begin running
Do we know if that is on a per-repository basis, or organization?
still zephyr
Could I run the Weblate CI on the core side? I have the ability just checking..
fossil gorge
Great job everyone, by the way, on 77 PRs merged. That's just.. 🤯
onyx hinge
@still zephyr is there one that is waiting due to that? You can feel free to press the button after you've made the most cursory check for the pull request being "real" (the "unreal" ones modify the github actions yml files)
idle owl
About this
> GitHub made it so first time contributors require someone to “appro...
I've added it to my list for tomorrow, I'll check it out.
onyx hinge
@still zephyr I've tried to address the weblate problem by adding the weblate bot as a "contributor"
but didn't see how it turned out yet
still zephyr
<@!607375730610798634> is there one that is waiting due to that? You can feel f...
Will do, just checking with you folks.. I saw the issue that you put in the weblate side 🙂
onyx hinge
@tulip sleet truth
lapis hemlock
I am lurking (I might have a short question in the "in the weeds")
idle owl
😊 You're welcome!
modern wing
Good luck on the vaccine recovery -- use it as an excuse for a bunch of rest, R&R, and hot soup 🙂
mental nexus
Dual espresso scale? ☕?
errant grail
Dual espresso scale? ☕?
... an early Clue version. The latest is on a PyPortal.
mental nexus
Whoa cool. Have been thinking of getting a machine but have generally felt overwhelmed by all the info.
idle owl
🎉
gilded cradle
The data files are created in circuitpython-build-tools
fossil gorge
FoamyGuy's stream recording mentioned: https://www.youtube.com/watch?v=Mjz0vMw-Jl8&t=6s
and YouTube channel: https://www.youtube.com/user/Zobrombee/
lone axle
Nice, Thank you Melissa
idle owl
Oops, I was out of order alphabetically as well. Thank you whoever fixed that.
thorny jay
@onyx hinge Would love to merge the colour of a low-res thermal camera with the image from a normal camera... could be grey scaling the normal camera and some alpha blending of the thermal false colour. 🙂
onyx hinge
@thorny jay I had exactly the same thought (well not necessarily about the implementation)
lone axle
The data files are created in circuitpython-build-tools
Do you know how the project bundler finds files that are used in the project, but aren't a library? Like in this project: https://learn.adafruit.com/electronic-history-of-the-day-with-pyportal/download-project-files-from-github when you download the project bundle it includes on_this_day_bg.bmp which is used in the project, but isn't from a library.
gilded cradle
That's done on the learn guide code. I believe there is some kind of scanning library that finds them.
thorny jay
They are on the github of the learn guide.
lone axle
Ah, I see. Maybe I can find them by listing files from inside the project folder in the learn guide repo.
gilded cradle
possibly
still zephyr
Will do after the meeting, I ll move the conversation to the graphics team
ok Ill move it there
lone axle
Maybe we could make a "Discussion" Repo that exists just to allow us to talk on issues?
thorny jay
yes
modern wing
@slender iron John Park encountered this on one of his live streams.
thorny jay
That was discussed in your stream comment I believe.
mental nexus
I kind of like having “conversation” under an issue since they seem to be easier to find/discover. Of course issues “conversations” can head in different off topic directions. Not sure if discussions are as discoverable?
thorny jay
Yeah... John Park!
blissful pollen
have to run, i'll catch the rest later
tidal kiln
^^ yep. just found same. and only thing that showed up in a recursive grep.
lapis hemlock
@tulip sleet Thanks for chiming in on github!
gilded cradle
I need to drop out a little early
still zephyr
@mental nexus we could discuss later, if you want regarding the mock imports.. on sphinx.
lone axle
I can help with moving these repos. I've gotten a start on some of the widget ones already, so I've got some experience with it already.
trim elm
@still zephyr I know hug reports are over but I wanted to give you a huge hug report for all those documentation updates! I've been out for the last week and I'm pretty sure half the emails in my inbox were from your documentation PRs lol. Thank you so much for doing all of that! Documentation is sooooo important especially for new users.
still zephyr
<@!607375730610798634> I know hug reports are over but I wanted to give you a hu...
🙂
mental nexus
Cool. Sounds like requirements.txt is the way to go if it can really “verify” the library imports functions really exist. Let me give it a shot on a library and see how it goes. Will using requirements.txt cause failures of “local” docs builds versus on GitHub?
still zephyr
I will try to put some guidance in the design guide after all my work in the documentation related to the sphinx built process in order for folks to understand better this process..
idle owl
@still zephyr Do you have a little time after this meeting? There's a process we usually follow when updating libraries, which includes merging the PR, and then doing a release. I'd like to have Dylan teach you how to do releases so you can take control of the entire process.
still zephyr
<@!607375730610798634> Do you have a little time after this meeting? There's a p...
I have yes
still zephyr
Cool. Sounds like requirements.txt is the way to go if it can really “verify” t...
No, for local you use mock imports
idle owl
@still zephyr Great. @trim elm - I'll ping you when this meeting is over and you can work with Jose David. Thanks to both of you!
still zephyr
Cool. Sounds like requirements.txt is the way to go if it can really “verify” t...
Ping me after the meeting, if you need help for sure
modern wing
Thanks everyone!
lapis hemlock
@onyx hinge Someone was asking about image processing functions, so if you have any suggestions, I am all ears.
lapis hemlock
The person wanted to have labelling and blob finding.
onyx hinge
I think you've already worked with the openmv people, yeah? we may want similar stuff to them.
still zephyr
Thanks
onyx hinge
Have a good week everybody! I'm stepping away.
mental nexus
Thanks all!
thorny jay
I don't know yet how Jeff camera work... but does the camera put thing into a "buffer" that can be processed by ulab?
idle owl
@trim elm The meeting is over. Feel free to work with @still zephyr anytime.
trim elm
@still zephyr Want to do a vc or whereby? It might be easier to explain stuff that way
lapis hemlock
I don't know yet how Jeff camera work... but does the camera put thing into a "b...
I am sure it supports the buffer protocol, so interfacing to ulab should be possible.
still zephyr
<@!607375730610798634> Want to do a vc or whereby? It might be easier to explain...
yes for sure..
mental nexus
No, for local you use mock imports
Ok will have to figure out if it goes in both places or only one. On the widgets I noticed that Melissa pulled out the mock imports and then added them to requirements. I don’t have a library right now to push. When I get to the Animation library I’ll give it a try and ping you with lots of questions!
idle owl
<@!607375730610798634> Want to do a vc or whereby? It might be easier to explain...
I suggest using one of the Discord voice channels. It will let others sit in if they wanted to.
There's the three "shared" ones.
trim elm
Ok. Let's go with Amelia Earhart
idle owl
And then you can also do text chat here.
onyx hinge
@lapis hemlock @thorny jay yup, it will work with anything that is a WriteableBuffer, which includes ulab arrays & displayio bitmaps. The "image format" is RGB565 so it is inconvenient to process (or it can be switched to YUV, which at least has the advantage that ever other byte can be treated as a greyscale value)
idle owl
Thanks!
crimson ferry
have the libraries diverged yet for 7.x?
thorny jay
I am sure it supports the buffer protocol, so interfacing to `ulab` should be po...
Because if that's a buffer that contain grey scale and I can multiply by a colour from the thermal... then we have the fake colour thing.
still zephyr
Ok will have to figure out if it goes in both places or only one. On the widgets...
No problem 🙂
idle owl
@slender iron I waved. Like it matters 😄
Let me know when you have a PR into the newsletter.
thorny jay
<@!676859171315908639> <@!681528999520108564> yup, it will work with anything th...
Seems that I will have to acquire some hardware...
slender iron
will do, just started the upload
lapis hemlock
<@!676859171315908639> <@!681528999520108564> yup, it will work with anything th...
I can wrap up the stuff here: https://github.com/v923z/micropython-ulab/pull/327
onyx hinge
@lapis hemlock oh will that help with pixel formats?
if so, excellent. I had not made a connection
thorny jay
@errant grail are you still around? I think you should really have a look at my thermal camera, I generated 64 colours palette in various way, made conversion from temperature to 64 values using ulab, used scaled bitmap for fast display. So some trick can be reused.
onyx hinge
@lapis hemlock CircuitPython 7
so not particularly quick
I'm very much in an investigative phase right now
lapis hemlock
@onyx hinge I have put that PR on the back burner, because there was no response from openmv, so I am not even sure at the moment, whether the implementation would fulfill the requirements. But if you have any comments on the general idea, I would really appreciate it.
onyx hinge
@lapis hemlock OK if I have something intelligent to say I'll add it
lapis hemlock
Thanks!
I am logging off now, but it was great to be around. Hope to do this more frequently.
errant grail
<@!339922377746022402> are you still around? I think you should really have a lo...
Yes, still here. I've looked at your code and use of ulab. My initial approach was similar, albeit with a very limited set of temperature values.
thorny jay
I believe palette colour are "free" so 64 or 128 are possible at no cost (once you started).
slender iron
thorny jay
Yes, still here. I've looked at your code and use of ulab. My initial approach w...
There as that FancyLED library that had function to generate palette, so at one point I tried that.
errant grail
@thorny jay In my current implementation, the colors are only limited by 24-bit RGB; no limited palette. Want to jump into a shared voice channel?
thorny jay
Marie Curie?
errant grail
See you there
still zephyr
@mental nexus I am now free if you want to discuss
idle owl
@tulip sleet Is there a way to view the frozen libraries from the REPL? Like importing sys or something? I thought for a moment maybe there was, but then realised I didn't know what I was doing.
mental nexus
<@!366059101203202052> I am now free if you want to discuss
Thanks for pinging me. But Sorry but I have to run. Let me get to trying to push a library and then that will help it make more sense and I will know the right questions to ask.
still zephyr
Thanks for pinging me. But Sorry but I have to run. Let me get to trying to push...
No problem 🙂
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/1z62nKbjNNv0j6uPWeUG6Co3aSCw4OQ3wl9VYlRIeTi8/edit?usp=sharing
still zephyr
Regarding the Graphics Widgets controls. I transferred the comments here https://github.com/adafruit/Adafruit_CircuitPython_DisplayIO_Layout/issues/19 to open the discussion on how we want to handle widgets behavior , if you would like to participate and leave your comments, we are in the process of moving the libraries and implementing the control functions. Any feedback is appreciated, testing, bugs, the normal stuff. Thanks.
manic glacierBOT
idle owl
Ugh, ok, I still can't figure out how to find the build assets off of a CircuitPython PR.
I see where they're being built, but.... wait... nevermind.
Artifacts.
🤦
Anyway, hurrah, the fix worked.
idle owl
@lone axle @slender iron I added the PyPI global secrets to the CircuitPython org on GitHub for the new circuitpython account.
slender iron
thanks!
idle owl
@slender iron For your own awareness, the email address attached to it is a new address (the mailing lists don't support +foo@ for addresses) that goes to circuitpython(at)adafruit.com so it's all the same folks.
slender iron
👍 thanks for sorting that out
idle owl
Yep!
lone axle
<@!382939733107408897> <@!252717193496756235> I added the PyPI global secrets to...
Was there one needed for readthedocs too? I'm not sure exactly how all of it gets hooked together.
idle owl
Was there one needed for readthedocs too? I'm not sure exactly how all of it get...
No, RTD is webhooks. @slender iron Do we want a separate RTD account as well?
slender iron
ya, I think it'd be good to have a shared RTD account like adabot
idle owl
On it.
idle owl
@lone axle Ok, so here's how RTD will work. You will generate the documentation on your own account. And then you'll add circuitpython as a "Maintainer" in Settings. That way, we'll all have access to the documentation in the event that we need to. We'll share credentials for that account as needed. This is how we do it for current docs, except with a different account.
still zephyr
<@!382939733107408897> Ok, so here's how RTD will work. You will generate the do...
I have added circuitpython to two of my libraries could you checked if that worked? please:)
lone axle
Sounds good. I'll try to get that set up on the first few repos that are in the circuitpython org this week.
idle owl
I have added circuitpython to two of my libraries could you checked if that work...
It did! Thank you!
idle owl
lone axle
would the URLs used in the readme need to point to my account? or by virtue of being shared to a maintainer do they work with the circuitpython in the URL?
still zephyr
the badge needs to be modified in the rst file
still zephyr
I have not move yet the docs, so in my case I am talking
idle owl
I think it should point to the shared account. So circuitpython. That way if you, for some reason, can't maintain things anymore, folks can still get to the docs. Does that make sense?
That isn't quite how we do it for the main docs. We have a CircuitPython project, and they're all added a subprojects, so the docs links are all the same.
slender iron
@tulip sleet do we want a milestone to track what fixes we want in a new 6.x release? most of them are going into main I think
idle owl
@fossil gorge OK, for the CI approval thing for first-time contributors, it appears to be a per-repository basis.
manic glacierBOT
still zephyr
@lone axle I added you as a Maintainer also in two repos, if you could verify that.. I could add you to the rest.
idle owl
@slender iron Maintainers with write access can approve CI runs for first-time contributors. So my assumption is that any role that has write access can approve.
slender iron
👍
idle owl
I think for now we see if the group we have keeps up with things. If we're struggling and missing PRs, we could look into adding more folks to the team, or adding folks as collaborators onto popular repos, etc.
lone axle
<@!382939733107408897> I added you as a Maintainer also in two repos, if you co...
Yep, I see these ones under my account now. Thank you!
still zephyr
Great thank you
manic glacierBOT
crimson ferry
@onyx hinge can you please point me to an example or ref on doing enums internally? re: https://github.com/adafruit/circuitpython/pull/4650#discussion_r620667708
manic glacierBOT
onyx hinge
@crimson ferry check out shared-bindings/displayio/__init__.c and .h
crimson ferry
thanks!
crimson ferry
@onyx hinge I'm still baffled. That doesn't look different to me than what I did. digitalio makes a class, which I've never done but presumably I could figure out. But your method is better than that I'm told, I'm just not getting it.
onyx hinge
@crimson ferry yeah the documentation should be better. Can I walk you through it another time though?
onyx hinge
OK, is there a good time for you tomorrow?
crimson ferry
I'm tied up until maybe mid-morning, then anytime
orchid basin
idle owl
@tulip sleet You around?
tulip sleet
yup
idle owl
What generates this? https://github.com/adafruit/circuitpython-org/blob/master/_data/files.json I see the it updates automatically with a release with a PR from adafruit_blinka, but I'm unclear on how, specifically, it knows what modules are built into a particular board. Where is it getting that infromation?
idle owl
Ah so it's in the core.
Ok, would it be possible to use Python code to generate a list of the frozen libraries for a board and include that?
Or some code.
idle owl
Also, from earlier, is there any way to tell what libraries are frozen into a board from the REPL? other than importing them. As in, is there a way to list them.
tulip sleet
i don't think there is
idle owl
Ok
tulip sleet
unless we can list the fake .frozen directory, let me get a CPX
idle owl
Thank you
I started to try, but realised I had no idea what I was doing.
Unrelated side note - this is excellent licensing, hadn't considered something like this: # SPDX-FileCopyrightText: 2014 MicroPython & CircuitPython contributors (https://github.com/adafruit/circuitpython/graphs/contributors)
tulip sleet
no, you can't list the fake .frozen directory
idle owl
I'm glad I bailed quickly then.
tulip sleet
but the frozen modules are included in help('modules'); you just can't tell which are frozen and which are native
tulip sleet
>>> help('modules')
__main__ adafruit_hid/keycode math supervisor
_pixelbuf adafruit_hid/mouse microcontroller sys
adafruit_hid/__init__ array micropython time
adafruit_hid/consumer_control board neopixel touchio
adafruit_hid/consumer_control_code builtins neopixel_write usb_hid
adafruit_hid/gamepad collections os usb_midi
adafruit_hid/keyboard digitalio random
adafruit_hid/keyboard_layout_us gc struct
Plus any modules on the filesystem
this is on a neopixel trinkey
idle owl
Ah yeah, that's what I have here.
The lack of differentiation is frustrating.
So as a general concept, is it possible to have this https://github.com/adafruit/circuitpython/blob/main/tools/build_board_info.py also include checking for frozen libs and generating a list where applicable? I'm unclear on what would go into that, but I'm sure there's someone else who could sort it out.
tulip sleet
we just have to find the frozen modules in mpconfigboard.mk, and add them to the download info
tulip sleet
we have to actually look inside the frozen/ directory, because only the repo names are in mpconfigboard.mk. So you'd see Adafruit_CircuitPython_HID, but not adafruit_hid in that file
idle owl
Ah hmm.
tulip sleet
this build_board_info.py script could do it
idle owl
Ok.
I'll file an issue on circuitpython.org and the core then and link them.
Thank you for your help!
tulip sleet
yes, I don't think this would be too hard.
crimson ferry
in RTD, sometimes there's a "Parameters" header for a function, sometimes there's not. Is there a distinction?
still zephyr
in RTD, sometimes there's a "Parameters" header for a function, sometimes there'...
Are you talking about c?
crimson ferry
both really, on the RTD website and in the code comments that generate it
classes typically call out parameters, but functions only sometimes have a Parameters header, otherwise they're just listed with a description of what they do, I'll find examples
crimson ferry
Maybe it's just variations of style, like code. Return values are similar... sometimes a "Returns" header, sometimes a "Returns" header and a "Return type" header, sometimes the return object is just in the descriptive text. Maybe there's a guide somewhere about this stuff.
prisma salmon
What have y'all been doing with these NeoTrinkeys? This is an awesome piece of hardware, and I'd love to hear what others are doing with it.
manic glacierBOT
tulip sleet
@crimson ferry sometimes we omit the parameters list if it is really obvious. If there are multiple parameters, it would probably be better to add an explicit list
idle owl
Just merged our oldest library PR! Thanks to s-light for sticking with it for so long, including all of the updates necessary to keep the PR in line with the library! 🎉
@still zephyr I created an issue on that repo about the documentation and assigned you.
still zephyr
🙂
idle owl
The docs appear to be there, but the TOC on the side seems incorrect. Anyway, verify it's an issue, and if not, simply close the issue.
still zephyr
ok, regarding the note, in newwer libraries the note regarding the I2C is not there,
I mean the board.SDA, board SCL,
idle owl
Ah ok
still zephyr
should I included?
idle owl
I think consistency is probably ideal.
still zephyr
good will do
still zephyr
😮
idle owl
What do you mean "they just create the object without the comment"
In the library
(I saw your comment now)
still zephyr
meaning, before this documentation storm, they were already libraries using the board.I2C object, not many but some, and they do not have the note
idle owl
Ahhhhh ok. I think if we're going to include the note, we should include it everywhere.
Even if it was already in the library without the note.
So, yes, continue as suggested before I said "hold on". 😄
still zephyr
ok, need to go back and verify, but no problem with that just a new PR
another question
idle owl
Sure
still zephyr
yesterday reviewing the libraries I saw this, do you happend to know where I can get one exactly like this?
idle owl
😊 Yeah... that's how they are in the shop unless the board has had a new revision. There's 3-4 different boards with my name on them. I designed the PCBs.
still zephyr
It is nice and unique
oh wow
I did not know that, just saw the picture yesterday
Thanks, I will go throught the rest of changes and then go back an do the note, is that ok?
idle owl
Thanks, I will go throught the rest of changes and then go back an do the note, ...
Absolutely! Whatever order works best for you. There's no rush.
still zephyr
Absolutely! Whatever order works best for you. There's no rush.
There is, summer is comming 🙂
thanks I will start with the ones this morning, thanks again 🙂
idle owl
You're entirely welcome! Thank you so much!
fossil gorge
😊 Yeah... that's how they are in the shop unless the board has had a new revisi...
As in hand-signed?
idle owl
As in hand-signed?
I mean, I didn't go to HQ with a gold sharpie 😆
idle owl
I created a bitmap and it was added to the silk library. It's part of the silk 🙂
idle owl
Exactly 😄
@gilded cradle Before I head down a stupid-hole, is it expected that my FunHouse does not have a UF2 bootloader on it? I'm doing all the things with reset and boot and it's not coming up as HOUSEBOOT.
It was one of the first run of them, I think.
gilded cradle
@idle owl it has a bootloader on it. Press reset and then again when the dotstars turn purple.
idle owl
🤦
@gilded cradle This is why I asked. Thank you. As a tangentially related side note, I can't download the funhouse_tinyuf2_combo.bin file in Chrome or Firefox, it tries to play it as a video.
gilded cradle
Oh, I think that's a learndev thing. Want to send them a note?
idle owl
Sure.
gilded cradle
thanks
They just added .bin downloading and must have made it a video type accidentally.
You can right click and choose save link as
gilded cradle
Yeah, except change it to .bin
gilded cradle
yeah, definitely
onyx hinge
@crimson ferry Feel like talking about enumerated values? I have time for more than a 2-sentence answer now.
crimson ferry
@onyx hinge Yes, thanks!
onyx hinge
@crimson ferry do you want to stick to text or do you prefer voice or video?
crimson ferry
text is fine, voice may be more efficient - I haven't done it in Disocrd but should work, I'm not set up for video
onyx hinge
OK, let me go to my desktop computer, I'll hear you in a few minutes in the Amelia voice channel.
crimson ferry
I think I have to restart Discord, back in 15s
analog bridge
Hi @onyx hinge and @crimson ferry... mind me listening in?
onyx hinge
@analog bridge of course not, though I think Anecdata is having technical difficulties so we may switch back to text
crimson ferry
@analog bridge no problem, but my audio isn't coming through I think
onyx hinge
silly discord
crimson ferry
sorry, I'm not sure what's wrong, permissions look ok
onyx hinge
that's OK, we can try it some other time
though maybe I'll screen-share in the "voice" channel, if it comes in handy
so.. I assume you're more interested in the "how do I use the enum stuff" than "how does the enum stuff work", right?
at least to start with
crimson ferry
This is the area of interest: https://github.com/adafruit/circuitpython/pull/4650#pullrequestreview-645137519
I can follow the PULL-like enum-like class I think, or @analog bridge's mesh, but Scott mentioned another way, but it wasn't clear to me looking at displayio yesterday
onyx hinge
OK I'll share my screen as I talk about what I did in displayio
I recently wanted to add color conversion in ColorConverter, and I used this new style of enum to do it
so the first thing is .. you need to include the header "py/enum.h" in the source files that work with the enums
where you get the enum as a parameter it is an MP_ARG_OBJ
and it can have a default value
are you seeing this in the screen share or do you want me to paste the bits of code in here?
analog bridge
stream is visible
onyx hinge
OK
crimson ferry
I can see it
onyx hinge
then when you use it to pass to the shared-module code, you have to use this formula where you call cp_enum_value (to turn a Python object to a C enum value), and cast it ...
it's a bit of a mouthful, but it does a lot of stuff, it checks that the Python object is the right type (so the value is a legal value), and throws an exception if it's not. So no need for explicit error handling
crimson ferry
yeah, I knew I had input validation issues
onyx hinge
so that I think covers how you'll get the value from a Python function call and into a shared-modules or common-hal function call. make sense? questions?
crimson ferry
knowing what to look at is a giant leap 😉
onyx hinge
the C enum gets declared in a header file, just like any normal enum
there are also some Python objects: The "type" object, and then in this case a specific one of the enumerated value objects, the one for RGB888.
That one's there because it was used as the default value. You can list them all if you like, or just the ones you need. sometimes you might not need any, but often you need a default one.
next, we get to the most magical part, the use of these C preprocessor macros. MAKE_ENUM_VALUE defines a Python object for a particular C enum value
so that first one defines the object for RGB888. The C name of the object is made up by pasting together the 2nd arg (displayio_colorspace), the 3rd arg (RGB888) and "_obj"
same as we saw in the header file's extern declaration
and then so on for all the other values. These don't have to come in order or anything, but it never hurts to be consistent.
I'll just call attention to the order of things when you put them in the docs -- the line for the documentation of a particular attribute comes right after it.
Next thing is to make the "map" (think Python dict) of the enum values. This is similar to what is done with the module globals, but again hidden behind macros
so you MAKE_ENUM_MAP, giving it that same prefix which is unique to your enum, and then MAKE_ENUM_MAP_ENTRY for each value
and then you have to MP_DEFINE_CONST_DICT and place it in your module globals ...
I skipped over these, but the first one creates the code that is necessary to print the enum objects in Python, and the second one creates the Python type object, the "class ColorSpace" object itself
so .. that's all the parts, I think
questions? comments?
crimson ferry
it;'s a lot to digest, I have to spend some time poring it over and trying it out... I was a little slow with my pen, can you point out the very first bit you talked about in the file?
analog bridge
nice! thanks @onyx hinge
onyx hinge
I'm not sure which file you mean, there was the MAKE_ENUM_VALUE
let me get you the link to this diff in github view and then you can link me to the lines you mean
crimson ferry
oh wow, yeah the diff will show me everything
onyx hinge
crimson ferry
OK, I think I have the scope of it, need time to digest the details, thank you!
onyx hinge
I'm happy to answer any questions and also take suggestions for documenting this for developers. that's absolutely not my strong suit
analog bridge
I can follow the `PULL`-like enum-like class I think, or <@!747673663419187281>'...
I think you can follow any of the methods... they all are different ways to achieve the same goal.
crimson ferry
i know nothing, I'll do whatever is best
onyx hinge
By the way, you can also take a look at the C header for enums https://github.com/adafruit/circuitpython/blob/main/py/enum.h and the implementation https://github.com/adafruit/circuitpython/blob/main/py/enum.c when you want to dive deeper
What I think is nice about this way of doing it (I'm biased, I created it) is that there's not much extra work to do to perform input validation and to print out the enum objects
One aspect I did not cover: if you have a C enum value and you want to return it to Python, you use mp_obj_t cp_enum_find(const mp_obj_type_t *type, int value); to turn it into the Python object...
https://github.com/adafruit/circuitpython/blob/HEAD/shared-bindings/canio/CAN.c#L179 so here we return a "bus state" value into Python from the C enum value
crimson ferry
good to know for future
Thanks for your time and help, sorry about the voice, I'll give this a shot.
onyx hinge
OK!
crimson ferry
anyone know why Discord wants to monitor my keyboard while I'm using other apps?
onyx hinge
@idle owl should I use "1.0.0" as the tag for an initial release of a library, or do we start with lower numbers like "0.1.0"?
crimson ferry
(trying to narrow down the audio thing)
onyx hinge
@crimson ferry It can use "push to talk" where (no matter what app you're in) holding a particular key will unmute your mic
(or to toggle mute)
idle owl
<@!330227457296957440> should I use "1.0.0" as the tag for an initial release of...
Depends? Is it a solid library? Or will there be a lot added to it immediately?
onyx hinge
@idle owl it's probably mostly done, the main caveat is you need the very latest CP for it to be of any use
😱
idle owl
Really it's entirely up to you, version numbers are free.
Hmm.
Maybe make it sub 1.0.0 then?
And 1.0.0 it when CP has a release that isn't S3?
onyx hinge
should there be a big warning at the top of the docs?
idle owl
Yes.
idle owl
Sure
onyx hinge
how better to word "unreleased development version"? "absolute latest" is what cp.o says I guess
idle owl
Put (absolute latest) in parens.
Because I would still think "latest" might mean stable if I was new.
onyx hinge
"Absolute newest"
idle owl
I guess whatever .org says is best
Since that's what should make people want to download it is the wording.
Did we update CircuitPython to do the new things with the status LEDs yet? Or is that still in the works. I don't see a PR for it, but my search fu might be failing.
onyx hinge
I don't think it has gone in
manic glacierBOT
idle owl
@gilded cradle So FunHouse doesn't really have a CircuitPython status LED then? Bootloader and reset, yes, but CP, no. I'm writing broken code and nothing is blinking. I wanted to make sure that was accurate.
gilded cradle
Oh, I'm not sure @idle owl. I usually have a serial console open so I didn't pay attention to that.
idle owl
Trying to make sure the wording I use on this guide page is accurate, is all.
I think I need to remove the word "status" in reference to "LED" in general because not all boards seem to use their LEDs as status.
gilded cradle
Let me try adding an error on my other funhouse that's not in sue and see what happens
idle owl
I was checking to see if it used one like Neo Trinkey, or multiple like others, or none.
gilded cradle
I would think a dotstar would light up, but maybe not
idle owl
Right. That was my thought. But CPX, for example, doesn't do status blinking. So it's not out of the question.
gilded cradle
Yeah, my lights don't blink either. Good observation.
idle owl
Ok, thanks for the sanity check. Appreciate it!
gilded cradle
yw
solar whale
@onyx hinge FYI -- I just adapted the new gcm4 ov7670 example to run with my 2.4inch TFT (ILI9341). It is running OK now -- with lastest CP build. I had some issues getting it running at first - board would hang but after a few power cycles and reloads, it is running OK.. not sure what the issue was.
slender iron
Did we update CircuitPython to do the new things with the status LEDs yet? Or is...
nope, I got distracted with merging
jaunty juniper
@idle owl note that the MagTag does light its top pixels purple for bootloader mode, but not yellow for safe mode (that seems to be common to all ESP32S2) or for errors/REPL. I wish it was more uniform and all boards that can did the purple (on ESP) and yellow light (including CPX) and just omit the rest on power-conscious boards like the MagTag does
onyx hinge
I was just noticing earlier that my qtpy m0 doesn't switch its LED on for safe mode (brownout)
@solar whale I see things like that too, and don't understand, haven't been able to find the pattern 😕
solar whale
Glad it’s not just me 😉
idle owl
<@!330227457296957440> note that the MagTag does light its top pixels purple for...
Thanks for the details.
idle owl
nope, I got distracted with merging
I was mostly making sure that it was that FunHouse didn't do status, not that CP had been updated to do the new status stuff yet.
slender iron
1.11 is merged
manic glacierBOT
crimson ferry
@onyx hinge sometime if you have a minute, could you take a look? I think I doted all the "i"s I knew of, but TypeError: can't convert Authmode to int when I pass one of the new enum values into a function
onyx hinge
@crimson ferry I think this needs to be a MP_ARG_OBJ now: { MP_QSTR_authmode, MP_ARG_KW_ONLY | MP_ARG_INT, {.u_int = WIFI_RADIO_AUTHMODE_WPA_WPA2_PSK} },
then you can say wifi_radio_authmode_type authmode = (wifi_radio_authmode_type)mp_enum_value(wifi_radio_authmode, args[ARG_authmode].u_obj or so
crimson ferry
I just found and fixing two old displayio refs too o_O
onyx hinge
oh? nice.
crimson ferry
(cut-paste)
onyx hinge
thank you!
crimson ferry
no, I meant in mine.. pasted and forgot to edit
onyx hinge
and the default in the allowed_args[] probably need to be .u_obj = wifi_radio_authmode_WPA_WPA2_PSK_obj
crimson ferry
i had tried that but it was still MP_ARG_INT and the compiler really didn't like that lol
still zephyr
Are we going to change the way we handle I2C, and SPI when MP is merged or it is going to be the same?
idle owl
@still zephyr Still on for chatting in a few minutes?
still zephyr
yes 🙂
idle owl
Excellent. I'll ping you in a little bit.
still zephyr
🙂
onyx hinge
@still zephyr i2c/spi won't change!
still zephyr
Thanks Jeff
idle owl
Ok, @still zephyr ready. Audio chat?
still zephyr
yes I am ready
idle owl
CircuitPython voice channel will work.
solar whale
WooHoo! Tried out the latest CP with MP1.11 on a feather_rp2040 and gcm4 -- both working well so far. @onyx hinge the ov7670 code still runs!!
jaunty juniper
I get AttributeError: 'PixelBuf' object has no attribute '_transmit' when trying to use a neopixel in the latest (QT PY M0 Haxpress specifically)
crimson ferry
also Saola, seems to be a NeoPixel issue
is there a button in the Github web GUI to un-commit the latest commit?
idle owl
[DotStar does not appear to be affected. (Relevant because both use pixelbuf.)]
blissful pollen
is there a button in the Github web GUI to un-commit the latest commit?
In a PR or a merge? I don't think there is a button to do it. I've done it but had to look up how as I'm far from a git expert
blissful pollen
You could just make another commit to undo the last one, not always ideal but may be fastest. I'm not sure if you can roll it back in the UI
crimson ferry
the git commands look risky too, I may do what you suggested
blissful pollen
If you're stuck on it let me know I can try to look up the commands / test them here.
ionic elk
@tulip sleet Jun2Sak did some cleanup - my testing shows the PR is ready for merge again. Should I encourage him to follow up with more cleanup in different PRs? The code works well, it just has a little bit of preprocessor junk in there that he or I could take care of later.
slender iron
<@!329766224093249548> Jun2Sak did some cleanup - my testing shows the PR is rea...
let's get it merged in
idle owl
One of the documentation PRs that Jose David put in was waiting on the CI run, so apparently permissions are irrelevant if it's your first PR to a repo. Pretty sure Jose David could have approved it, but still, it was a question we were unsure of.
idle owl
@still zephyr Sorry for delayed review comments. I think that should be it.
Has anyone seen anything like this before? I'm ready to replace it, but I'm not sure what would be causing it, and want to make sure it's not going to end with another replacement. https://forums.adafruit.com/viewtopic.php?f=60&t=178640
tulip sleet
<@!329766224093249548> Jun2Sak did some cleanup - my testing shows the PR is rea...
Why don't you just clean it up yourself; you should be able to push a commit to his branch. or submit another PR. If you're happy with it, let's merge it.
still zephyr
<@!607375730610798634> Sorry for delayed review comments. I think that should be...
No problem, I was trying to figure out how to stop the CI 🙂
idle owl
Has anyone seen anything like this before? I'm ready to replace it, but I'm not ...
If anyone has a moment to take a look at this and let me know if it's ringing any bells, I'd appreciate it.
blissful pollen
If anyone has a moment to take a look at this and let me know if it's ringing an...
I looked at it, and never seen anything like that before but I'm not that knowledge with bootloader stuff and why it wouldn't hold it after a power reset
idle owl
I looked at it, and never seen anything like that before but I'm not that knowle...
Thank you for taking a look.
ornate breach
So a small typo change i'm including in my PR is something i've been meaning to fix for a few months but it's completely got buried in the depths of my mind.
manic glacierBOT
ornate breach
dang, thought i was going to get that fixes to devices.h pushed in that PR but there's a merge conflict on it. next PR it is.
rose dove
so i've been messing with the wio terminal for a bit trying to get neworking going, and I ran into some troubles that I was hoping I could get some advice on
specifically, the board is missing microcontroller.pins.PB24. I added this into atmel-samd/common-hal/microcontroller/__init__.c, and it shows up now, but when I try and use the pin for a uart, I get a hard fault. I suppose I'm missing some magic wrt adding pins, and was hoping someone could point me in the right direction
meager fog
@lone axle hay any luck with the pillow haxing
lone axle
Heya, yes. I have gotten it to generate for all of the projects in learn guide repo!
it revealed some oddities with a few of projects that I am chasing down now actually
lone axle
I found a few files with UTF BOM in them and findimports raises an exception from it.
meager fog
UTF BOM??
lone axle
Early hug report: Would have taken me ages to figure out without Hugo's help!
Here are the screenshots generated by the current iteration of the script: https://github.com/FoamyGuy/CircuitPython_Library_Screenshot_Maker/tree/main/python_generator/generated_images
Still have a few things to clean up with them. But well on our way to getting them all generating nicely.
ornate breach
nice job @fossil gorge
fossil gorge
You know what the kids say... BOM me once, shame on you, BOM me twice, then..., you... we won't be BOM'd again
ornate breach
i didn't know they said that, but it makes sense 😉
jaunty juniper
BBEdit offers the "unicode BOM" and "unicode NO BOM" encodings, I've never seen it mentioned anywhere else
fossil gorge
We found that there's a "Remove BOM" in Pycharm. But if you don't know to look for a BOM, that well-hidden option doesn't help
For the record: File -> File Properties -> (several options)
meager fog
@lone axle looks like its close!
lone axle
Yep, have some spacing issues on some it seems and need to make sure none of them have unnecessary files included.
ornate breach
hey @tulip sleet if i had to delete my fork and refork, does it then count me as a new contributor?
Or is it the amount of time since my last PR?
tulip sleet
hey <@!329766224093249548> if i had to delete my fork and refork, does it then c...
are you asking about the PR build being held up?
tulip sleet
it's a new feature at GitHub, they haven't tuned it a lot yet. I think everyone is getting held up the first time right now.
ornate breach
ah okay makes sense
tulip sleet
approved to run
ornate breach
That PR is mostly just waiting on my PR for PID Codes to get merged. (assuming no one took 4DDF)
well, looks like i have the only PR for PID codes so it doesn't look like i'll have any issues with that.
ornate breach
weird.. ja build for my board failed due to too little flash
hollow token
Is this the right channel to ask about library PRs? or is this only for the main CircuitPython repo?
ornate breach
this would be the right place
hollow token
I did ask in #help-with-circuitpython yesterday but I didn't get a response and I don't know if I am being a bit eager with these changes, haha
ornate breach
if it's an Adafruit library and you want someone to take a look you can ask here
sometimes things get buried by github notification though
hollow token
Sweet, I'm looking to add some extra chips to the MCP230XX library, I've got a PR here, https://github.com/adafruit/Adafruit_CircuitPython_MCP230xx/pull/41
ornate breach
okay so it looks like you latched circuitpython librarians for a review of it
i would just give it a little time
hollow token
ok
ornate breach
it can usually take a day or two as they are working on issues across a number of different libraries, mostly circuitpython itself
hollow token
Thanks!
ornate breach
🙂
hollow token
@still zephyr @idle owl I'm right here if you'd like to talk about https://github.com/adafruit/Adafruit_CircuitPython_MCP230xx/pull/41
analog bridge
is there a button in the Github web GUI to un-commit the latest commit?
@crimson ferry assuming you are taking about PR #4650... you can remove changes after c472245 by:
git reset --hard c472245
git push -f
@analog bridge The current state should be exactly that. I just did it manually.
analog bridge
suggestion needed regarding wifi authmode property, is it better to have combined enum field like WPA_WPA2_PSK or an array of separate like [WPA,WPA2,PSK].
the latter has the advantage of less QSTR
manic glacierBOT
idle owl
<@!607375730610798634> <@!330227457296957440> I'm right here if you'd like to ta...
Thank you for the PR! It's best to leave discussion of a PR in the conversation thread of the PR. I appreciate your excitement, but as skerr said, it can sometimes take a bit for a merge to happen, especially when it's a more complicated addition that requires testing (versus, say, a documentation fix). I see that there was already follow-up, so I'll leave it to the already-involved folks for now.
@tulip sleet You're the only other person I can think of for whom it's worth taking a look at this: https://forums.adafruit.com/viewtopic.php?f=60&t=178640 Have you seen this behavior before? I'm prepared to replace it, but I don't want it to turn into repeated replacements if it's something within their setup.
hollow token
Thanks for replying back :)
idle owl
You're welcome! Thank you for the contribution! Keep us posted in the thread with how testing goes.
hollow token
Will do
tulip sleet
Thank you for the PR! It's best to leave discussion of a PR in the conversation ...
That is bizarre. I would make sure they are trying it with no code.py. A couple of q's: Are the files in CIRCUITPY gone after the reload of the cpy .uf2, or are they stil there? I would suggest reloading, leaving the power on, and doing a storage.erase_filesystem() and see what happens, including after unplug/replug.
If the CIRCUITPY filesystem gets reset on each reload, then there may be something broken about the flash.
idle owl
Ok, thank you.
tulip sleet
i.e. I am wondering if something in code.py is bringing it down
idle owl
Got it. I'll ask.
@tulip sleet Alright I replied and asked the questions, and asked them to do erase_filesystem(). I'll let you know what comes of it.
tulip sleet
I'll subscribe to the topic.
idle owl
That also works.
🤦 I did an erase_filesystem() to make sure I described the process properly. Then dumped code back on my board, and wondered why it wasn't running. Looked at the serial output on the tiny display and saw it said something about adafruit_dotstar and thought, oh I must have misspelled it, it doesn't look misspelled though.... no, Kattni. There's no libraries on the board anymore because you erased them.
@tulip sleet Wait, have we added the thing that will break the 6.x .mpy lib files to CircuitPython yet?
I'm running the latest of main as of yesterday as Limor requested. Libs are still working, so I guess not?
tulip sleet
there are built mpy-cross's in https://adafruit-circuit-python.s3.amazonaws.com/index.html?prefix=bin/mpy-cross/
we build one on every pr
idle owl
Oh snap.
Perfection.
Thanks @tulip sleet! I would have fought with building it for no reason 😄
tidal kiln
any reason the Matrix Portal page doesn't have a UF2 bootloader section?
https://circuitpython.org/board/matrixportal_m4/
ionic elk
@tulip sleet @slender iron I agree, I'd rather just merge it in now. Maybe we can follow up with an issue for the NRF stuff @analog bridge noted was out of place.
@tulip sleet I believe it's just waiting on your review approval/dismissal
tulip sleet
@ionic elk have you tested then NRF PR pretty thoroughly, both real and simulated deep and light sleep?
ionic elk
@tulip sleet Yeah, I've tested fake/real light/deep, it all works fine.
The only thing that's weird is that it doesn't keep a timer through real deep sleep, so it does the USB delay every time
tulip sleet
Then I will just dismiss my review. I don't want to go through it again. Could you open an issue about that?
ionic elk
Not a huge deal for long deep sleeps, but something to fix later I think
Sure
I'll open issues for anything remaining
tulip sleet
@ionic elk can you also look at the "hidden" comments, and check them all and resolve them to make sure they have been addressed? That is necessary to unblock the merge.
ionic elk
@tulip sleet ok. @analog bridge's serial.c issue is not resolved but I'll open an issue for it
analog bridge
@ionic elk #4680 covers that
tulip sleet
is that #4680 above I just opened
ionic elk
Oh sorry I missed that
I kinda glaze over the github bot sometimes
I think we're set, if you're able to dismiss now
tulip sleet
@ionic elk bang. merged.
idle owl
@tulip sleet They replied. Not the response I was hoping for, which is to say, I was kind of hoping it would fail because that's an easier situation than troubleshooting this weirdness.
ionic elk
@tulip sleet Great, thank you! I will clean up the STM32 PR given the new changes, should be done today pending no unexpected disasters
tulip sleet
<@!329766224093249548> They replied. Not the response I was hoping for, which is...
They could try another machine, like a Mac or RPi. The fact that Windows doesn't see it at all when it's plugged back is really weird. Otherwise I'd say just replace, it's cheap, and our time is worth something.
idle owl
They could try another machine, like a Mac or RPi. The fact that Windows doesn't...
Agreed. I'll ask about another machine, and then replace it regardless. Thanks!
analog bridge
suggestion needed regarding wifi authmode property, is it better to have combine...
@slender iron ^^
manic glacierBOT
idle owl
@gilded cradle Folks mentioned that _transmit error yesterday on QT Py M0 Haxpress and Saola.
gilded cradle
Thanks @idle owl. I noticed it on the FunHouse.
idle owl
Thanks <@330227457296957440>. I noticed it on the FunHouse.
Simply wanted to make sure you knew it wasn't FunHouse-specific.
gilded cradle
Yes, that's helpful. I'll make a note on the issue.
manic glacierBOT
slender iron
<@!252717193496756235> ^^
I'd do an array or bitmask
combinatoric enum will just get bigger and bigger
solar whale
@onyx hinge that was a lot of wires! I now have the ov7670 working on the pico! Thanks for the examples!
still zephyr
<@!607375730610798634> <@!330227457296957440> I'm right here if you'd like to ta...
🙂 You found me 😆
slender iron
@onyx hinge when you have a chance I could use your help fixing an re test
onyx hinge
@slender iron sure, what's up?
still zephyr
@idle owl Question, During the weekend I want to go back and revise that we have the note regarding the pins in board.I2C, Do you want me to leave you all the approvals for monday (if any) or do you want me me to auto-approve?
onyx hinge
3 tests failed: builtin_override qstr_limit ure1 specifically looking at the ure1 failure?
slender iron
I tried looking at it but don't know how to fix it
ya, I have a fix for qstr_limit already
idle owl
<@!330227457296957440> Question, During the weekend I want to go back and revise...
Always worth having another set of eyes. Please leave them for me.
slender iron
I have 6 tests failing. weird you only have 3
still zephyr
will do, by the way this morning I saw that indeed the community_bundle is aleady in there, does not know how.. 🙂 but is ok i guess
slender iron
did you do make coverage_test?
idle owl
will do, by the way this morning I saw that indeed the community_bundle is alead...
Right?! I have no idea either. But it's good to have.
still zephyr
Right?! I have no idea either. But it's good to have.
Yes, and it was me indeed, so good at least Scott was reviewing those changes at that time
idle owl
Like I said, we can always revert if necessary. But it worked out!
still zephyr
Yes, thanks Kattni
idle owl
@slender iron Do you know what Winbond chip ships on the Itsy RP2040? I can't read it and I need the dimensions. I figured you might know with your massive flash collection.
onyx hinge
@slender iron is it possible the change to extmod/re1.5/compilecode.c just got lost in the merge? ebf8332104c1
slender iron
<@!252717193496756235> Do you know what Winbond chip ships on the Itsy RP2040? I...
I think it's the XSON package
idle owl
Right on thank you
onyx hinge
no, that breaks it different/worse
slender iron
right, I think you changed our version in a different way
onyx hinge
I'll keep looking into it and let you know.
I was about ready for a concrete diversion anyway
slender iron
thanks!
idle owl
@tulip sleet Replacing that person's QT Py RP2040. Here's hoping it resolves the issue.
idle owl
I think it's the XSON package
The dimensions fit the silk perfectly, so yeah that's correct.
slender iron
👍
idle owl
Thank you for the help.
slender iron
np
onyx hinge
@slender iron OK I sent you a pull request but feel free to incorporate into the merge commit
remaining test failure here is 1 tests failed: builtin_override
onyx hinge
huh, not here. Is it intermittent or consistent for you?
slender iron
its very weird
it fails from run-test but the output looks ok when I run it directly
onyx hinge
oh that's a skip here
slender iron
what command are you running?
onyx hinge
"./run-tests" in the tests/ directory
slender iron
I'm running make coverage_test from the unix port dir
slender iron
yup yup
onyx hinge
3 tests failed: builtin_override mpy_native vfs_fat_fileio1
slender iron
k, that seems like where I'm at
onyx hinge
OK, I'm going to step away but I can work on one of them a bit later.
slender iron
(I have two other failures due to me comparing against 3.9)
the first and third may be related
tulip sleet
@slender iron sorry to interrupt, i have a q about using storage_allocations to save the state of some Python objects across VM instantiations. When the heap is torn down (cleanup_after_vm()), immediately after that, would it be safe to assume I can still copy heap objects out to a storage_allocation? (The point of this is approximately to save and restore usb_hid.devices between boot.py and code.py, though it's a tiny bit more complicated than that).
I originally tried to make a storage_allocation of the right size while the heap was alive, but that's not possible. I could always make a large-enough storage_allocation by limiting the number of devices, but if I don't have to do that, it would be preferable.
tulip sleet
the q is really just if after the heap it freed then its contents are still viable to copy or not
or if it might immediately get smashed by somethign
slender iron
ya, I think so. teardown doesn't blank. you just risk allocating something over it
tulip sleet
right, but maybe i can do it right away. I'll see if i feel comfortable about that, thanks
slender iron
👍
ionic elk
Did something just change with MP_OBJ_IS_TYPE?
My builds for NRF and STM32 use it in uppercase, and are suddenly failing... it's not a hard fix, but how did it get past CI on the NRF PR we just merged? I built and tested that just yesterday
blissful pollen
Did something just change with MP_OBJ_IS_TYPE?
I think the MP 1.11 update moved all the macros to lower case. That may have just been merged
ionic elk
So it was... is that just a CI thing? I guess CI ran before that was merged, and then it didn't re-run before we put NRF in
ionic elk
so there's just a little window of time where a PR can actually be failing CI but look like it's ok
slender iron
the fileio1 issue is related to fast seek but I'm not sure how
ya, some projects do require a latest merge CI pass but our CI is costly
so I think it's better to just fix it when it breaks
ionic elk
yeah that's ok by me just wanted to see if this was a broken thing or just a general risk
slender iron
that depends on what you are going for 🙂
I think for us, it's just a risk we're ok with
ionic elk
I'm going to hotfix NRF then so it's not broken
slender iron
kk, thanks
manic glacierBOT
onyx hinge
@slender iron I'm back if you want me to see if I can figure out the fileio1 thing..?
er wait is it passing now?
tulip sleet
@ionic elk I will merge as soon as it builds right
onyx hinge
after pulling from you I have 2 tests failed: builtin_override mpy_native
debugging mpy_native, it looks like it arrives in mp_native_relocate pointing 1 byte earlier, at b"\x04" # bss, 4 bytes instead of b"\x03\x01\x00" # dummy relocation of rodata, which causes a write through a NULL pointer
slender iron
@onyx hinge yup, I fixed fileio. I'm taking a break for lunch and a walk so feel free to look at the other two. 🙂
onyx hinge
@slender iron fix for mpy_native https://gist.github.com/feb8a0e129ca44937238a11ef9e734f2
builtin_override: Setting builtins __build_class__ back to its initial value is not quite working for some reason
idle owl
Wheeeeeeeee MpyError: Incompatible .mpy file. Please update all .mpy files. See http://adafru.it/mpy-update for more info.
Wait, what? -bash: ./mpy-cross: Permission denied Since when?
tulip sleet
just chmod +x
idle owl
Do I need to chmod the mpy-cross file? Or am I supposed to be running it as sudo?
Ok thanks
tulip sleet
the download does not carry those permission bits
idle owl
Oi.... I had to give it permission through System Preferences too. MacOS is NOT happy. Heh.
There it goes. Finally. Thanks Dan!
tulip sleet
our evil non-verified executables
idle owl
Pretty much.
At least I'm running into it now, so we know what to tell folks in the near future.
Rainbow!
(It worked.)
tulip sleet
is this a small board? Maybe in other cases you can get away with the original .py's
idle owl
Yeah I thought about that after I went through that whole process. It is the opposite of small. Itsy RP2040. Definitely could have just used the .py. 🙄
But, Limor wants me running latest for everything I do moving forward, so I'll need to mpy-cross things soon enough.
tulip sleet
i wonder what the breakpoint is recommending the py source bundles instead of the mpy bundles. mpy is always faster and less fragmentation, at the cost of more complexity
idle owl
Well when we're making the mpys, there's no reason not to tell folks to use them.
tulip sleet
but, we didn't start with RP2040 boards... Yes, I am just thinking about an alternate universe
idle owl
But when expecting folks to make their own mpys.... good question.
idle owl
but, we didn't start with RP2040 boards... Yes, I am just thinking about an alte...
Yeah, exactly.
I mean we had to start somewhere. M0 was it, and we can't simply ignore them now. 😄
As much as there are days I would love to.
still zephyr
@idle owl Question 🙂
idle owl
Answer!
still zephyr
jajaja
idle owl
What's up?
still zephyr
Did you validate the standard on how to showcase adafruit products in the libraries?
idle owl
Probably not. How do you mean?
still zephyr
there are a lot of different ways in the libraries, now that I am going back to verify the board note I would like to do also that
so for example
idle owl
Ah, right now they're all different is what you're saying.
And you want to know what's best.
still zephyr
* Adafruit `BME280 Temperature, Humidity and Barometric Pressure sensor
<https://www.adafruit.com/product/2652>`_ (Product ID: 2652)
* Adafruit `BME280 Temperature, Humidity and Barometric Pressure sensor
<https://www.adafruit.com/product/2652>`_
* `Adafruit BME280 Temperature, Humidity and Barometric Pressure sensor
<https://www.adafruit.com/product/2652>`_
Adafruit BME280 Temperature, Humidity and Barometric Pressure sensor
<https://www.adafruit.com/product/2652>
Ok, I'm unclear on what the backticks are doing.
still zephyr
All render different
idle owl
Right ok
still zephyr
the backstick will indicate where the hyperlink will start
still zephyr
The last one, does not have even so only the adress
idle owl
I know what the asterisk is doing at least.
still zephyr
the underscore will indicate the end of the hyperlink
idle owl
Ok.
still zephyr
I like the first one, but is not for me to decide
The second one is nice too
I can render all and put a picture here
idle owl
I think a combination of the first and third one? * `Adafruit BME280 Temperature, Humidity and Barometric Pressure sensor <https://www.adafruit.com/product/2652>`_ (Product ID: 2652)
I think the whole name should be a link, it is a little odd to leave out only the "Adafruit".
idle owl
I can render all and put a picture here
That would be great.
idle owl
Thanks!
idle owl
Ok, I think leave out the 's at the end of Adafruit, and the : (colon) at the end of the URL. And go with the first one.
idle owl
Does that make sense?
still zephyr
I will verify, give me 1 minute
still zephyr
Does that make sense?
idle owl
Yeah! I think that works. What do you think?
onyx hinge
@slender iron https://gist.github.com/82147a70aadf7fcab9ecc938cab9144a another fix for ya
now the first set of tests pass, but the threaded tests just failed...
1 tests failed: thread_exc2
make: *** [Makefile:261: coverage_test] Error 1
aha, that was trivial
slender iron
thanks @onyx hinge! I'll integrate it shortly
onyx hinge
@slender iron do you want PRs or are the gists okay?
idle owl
Read that as grits.
jaunty juniper
I didn't find any info on whether the old mpy are compatible with the new versions of circuitpython, I assume they are ? (1.11 switched to version 4 of mpys)
slender iron
gists are fine. I applied the first one by hand anyway
idle owl
I didn't find any info on whether the old mpy are compatible with the new versio...
Not anymore! Latest requires a new version of mpys.
Or at least it did for me just now.
jaunty juniper
ah ok then
onyx hinge
thread_exc2 fix: https://gist.github.com/3f8052071ee3187b6b141fbb3650f065
@slender iron ^
OK now make coverage_test completes successfully
cobalt grail
Some info on mpy versions - https://docs.micropython.org/en/latest/reference/mpyfiles.html
slender iron
@onyx hinge looking at your first patch, you set ASYNC to 0x00
onyx hinge
oooops
slender iron
which will break some of our async stuff
slender iron
ya, 1.13 has async stuff in it as well
onyx hinge
do you mind just manually correcting it?
onyx hinge
they do
anyway I pushed but didn't create a pull request from my branch (updated with ASYNC at 0x80): https://github.com/tannewt/circuitpython/compare/merge_1.12...jepler:1.12-more-fixes?expand=1
tests/import/mpy_native.py: b"\x70" # scope_flags: VIPERBSS | VIPERRODATA | VIPERRELOC
tools/mpy_ld.py:MP_SCOPE_FLAG_VIPERBSS = 0x40
``` so 3 times, one in C, one in py, and one implicitly in py@tulip sleet Gah, I guess I realised that about individual commits to the core, but I wasn't thinking about it. I will be sure to include that in any core reviews I do with multiple change requests. Sorry about that!
NeoPIxel works fine on latest for me. What were folks doing when it was failing with the _transmit error?
slender iron
@onyx hinge it looks like async_await fails when FLAG_ASYNC is 0x80
slender iron
I'm looking at why now
onyx hinge
It seems nicer if our flag ordering matches theirs
slender iron
ya
onyx hinge
../../py/emitglue.h: mp_uint_t scope_flags : 7; not enough bits
slender iron
ah
onyx hinge
mp_uint_t kind : 3; // of type mp_raw_code_kind_t
mp_uint_t scope_flags : 7;
mp_uint_t n_pos_args : 11;
const void *fun_data;
``` can increase it to 8 , there are currently 11 bits left out of 32"interesting" that up to 2**11 positional args are supported. why so many.
slender iron
¯_(ツ)_/¯
thoughts? ```c
uint16_t kind : 3; // of type mp_raw_code_kind_t
uint16_t n_pos_args : 11;
uint16_t reserved : 2;
uint8_t scope_flags;
could make scope flags 16 then too
jaunty juniper
NeoPIxel works fine on latest for me. What were folks doing when it was failing ...
just the basics, like
import board
import neopixel
pix = neopixel.NeoPixel(board.NEOPIXEL,1)
pix[0] = (0,255,0)
It seems to leave the struct the same size @slender iron so from that viewpoint it's fine
jaunty juniper
but right now, on QT PY Haxpress, latest (20210428-d4d96bb) gives me CircuitPython core code crashed hard. (fresh from erase_filesystem())
idle owl
just the basics, like
```py
import board
import neopixel
pix = neopixel.NeoPixel...
Odd. The only difference in my code is I did not specify a particular pixel.
slender iron
it'll simplify the loading code too
onyx hinge
you could eliminate the bitfields altogether in the same 4 bytes, couldn't you? uint8_t kind; uint8_t scope_flags; uint16_t n_pos_args;
slender iron
yup
idle owl
just the basics, like
```py
import board
import neopixel
pix = neopixel.NeoPixel...
This code makes it red, then green, then blue. I also ran rainbow code. No errors.
jaunty juniper
Odd. The only difference in my code is I did not specify a particular pixel.
oh yeah pix.fill() works on the version I had the _transmit error
(using the .py in both cases)
idle owl
oh yeah pix.fill() works on the version I had the _transmit error
Ok, specified a pixel, and got the error.
With a .mpy I generated a little bit ago from the .py.
Probably worth mentioning on the issue.
Wait a tick.
I do specify pixels in the rainbow code.
Curiouser and curiouser.
fossil gorge
our evil non-verified executables
Mining the bitcoins, one MPY at a time!
ionic elk
I'm having trouble totally understanding what's up with the mpy stuff - if I want Neopixel support while testing my branch off current main, how do I do it? Doesn't seem like the library file by itself works either
fossil gorge
But when expecting folks to make their own mpys.... good question.
This is something I've been thinking about for a bit - ever since Jeff E. showed off the "Convert a font to BDF" (I think? or maybe "from BDF") site.... Is something where a user could upload a .py and get a .mpy file something that might be useful? Or is downloading MPY-cross and running it simple enough?
idle owl
This is something I've been thinking about for a bit - ever since Jeff E. showed...
It's pretty simple. But it is a command-line thing, and not everyone is comfortable there. Fact is, once there's a 7 release, we'll be generating a .mpy bundle, so it's a non-issue. And folks that are comfortable living on the bleeding edge of development, should be ok with running mpy-cross if they need to.
idle owl
I say "should" not in a declaratory sense, but in the sense that they're probably ok with venturing into the command line even if they've not dealt with it before.
fossil gorge
I think a combination of the first and third one? ```* `Adafruit BME280 Temperat...
Sorry for the late intrusion, but one thing I was thinking is that it might be useful to use the adafru.it/<PID> URL shortener for product links. That way they would always end up on the product page if the URL structure should ever change.
What made me think of this is that those are the URLs that are on the individual product packages - though space there is at a premium
idle owl
Sorry for the late intrusion, but one thing I was thinking is that it might be u...
I genuinely had no idea that there were short URLs for products.
manic glacierBOT
fossil gorge
It's pretty simple. But it is a command-line thing, and not everyone is comforta...
That's very true - very good points. I was thinking more those who was running into space/memory constraints for their projects on the devices, but that might also be a "if size is that tight, this may not be the right board for the project". 🤔
idle owl
That's very true - very good points. I was thinking more those who was running i...
Well folks who are writing their own libraries and so on should definitely be able to run it. I mean, you download a file, and then for easiest use, put it in the directory with the file you want to mpy-cross, and type ./mpy-cross file.py (at least on MacOS). It's not difficult.
fossil gorge
Very true. Glad I asked, before I started a solution for a not-a-problem!
Thanks for the insight! 🙂
idle owl
You're entirely welcome! I have wasted plenty of time fixing things that shouldn't be.
fossil gorge
Those rabbit holes are really quite enticing!
ionic elk
@tulip sleet NRF hotfix CI is green
Also STM32 is re-tested and all set, no pressure on that though
slender iron
it looks like the latest mpy format is hex encoded bytes. do we want to make them proper binary files?
tulip sleet
it looks like the latest mpy format is hex encoded bytes. do we want to make the...
i'd think so. hex encoding for small filesystems, why?
tulip sleet
i wonder if it got changed back
slender iron
I think the version from 1.12 is the latest
@vale spire any idea about mpy not being binary? 👆
tulip sleet
@slender iron I don't see that. I did an mpy-cross built from the tip of micropython, and it produces binary
slender iron
oh, it could be my editor being fancy
false alarm @vale spire 🙂
@onyx hinge sound "prelude" encoding in bc.h
onyx hinge
oh that is complicated looking.
does that break the things we were talking about doing with the structure?
slender iron
I think if I do renumber viper I can have mpy include 5 bits
just gotta update the other spots
onyx hinge
@slender iron was skimming release notes and saw this: gc: add gc_sweep_all() function to run all remaining finalisers
Maybe we want to try that at vm shutdown
slender iron
👍
manic glacierBOT
@RWhite-EDR Actually you could starting know if you download the firmware built for your particular board in the PR
https://github.com/adafruit/circuitpython/pull/4650
Here
https://github.com/adafruit/circuitpython/pull/4650/checks
If you already knew ignore the previous message 😄
ornate breach
Hey @idle owl could you trigger CI for my PR? https://github.com/adafruit/circuitpython/pull/4677
i updated to try and see if I got the make command right for doing Ladyada's suggestion for the JA build failing due to size
mental nexus
Running now.
ornate breach
Thanks @mental nexus
bright imp
Hi, I have a question, I got a qt py and a TM 1637 7 segment display, got it all wired up. When I put the TM1637 file in the /lib folder, then in the code.py folder try to import it import TM1637 it does not work. I've narrowed it down to the import statement which leads me to believe it cannot find the file to import.
Referencing this github repo: https://github.com/bablokb/circuitpython-tm1637
ornate breach
hey @bright imp this channel is intended for Circuitpython development discussion, I saw you posted in #help-with-circuitpython
bright imp
hey <@!152975259233026048> this channel is intended for Circuitpython developmen...
you mean the language itself not working with it?
I just wasnt sure where the best place is to post, feel free to delete if you'd like
ornate breach
I can't delete, we just try to reduce cross posting
i don't think that library is maintained by Adafruit
did you get it from the community bundle?
bright imp
The lib im trying to import is from that github repo, but it seems to not be able to import anything. Even if I put a simple class in the .py file under /lib it wont import it
I can delete my post if you would like, lmk
ornate breach
Let’s take this back over to the help with circuitpython
bright imp
👍
crimson ferry
@ornate breach you need your workflow kickstarted? I seem to have a button for that
ornate breach
Yeah, I think I got it fixed now to pass
misty garnet
no mpy-cross 7.0 for windows binary?
manic glacierBOT
ornate breach
Something is still amiss. Hmm
solar whale
@tulip sleet got a minute for a bootloader question?
tulip sleet
sure, but if it's ESP32-S2, I'm no expert.
solar whale
may be more general -- I amnually loaded thetinyuf2 bootloader from the repo to a funhouse. Should I next copy over the update bootloader as well?
I'm still confused by the role of the update_bootlaoder?
tulip sleet
I'm not sure about the tinyuf2 update, but I would assume it's just to update the bootloader to a newer version. So if you've already loaded the latest, I wouldn't see a need to update it.
are there version numbers for each?
solar whale
ah -- ok so it just does the update, it does not modify the currently loaded bootloader
tulip sleet
I would think with the name "update_bootloader", it does update the bootloader