#circuitpython-dev
slender iron
Lost my internet @onyx hinge
onyx hinge
@slender iron we can surely finish without you π
mental nexus
I've got a custom CP build running on the ESP32-S3 devkit N8R8, and I can't seem to access my IO41/42 for I2C SCL/SDA. Whenever I do busio.I2C(board.IO41, board.IO42), I get an error: ValueError: IO41 in use.
I have pullups on both pins, so that can't be the issue. Also, the same happens with a hard reset or with the push button reset.
mental nexus
It's an older build. Here's boot.out:
Adafruit CircuitPython 5.3.0-1808-gfcde108d0-dirty on 2022-04-20; ESP32-S3-DevKitC-1-N8R8 with ESP32S3
Board ID:espressif_esp32s3_devkitc_1_n8r8
stuck elbow
I have pullups on both pins, so that can't be the issue. Also, the same happens ...
do you have anything in board.c?
mental nexus
do you have anything in board.c?
void board_init(void) {
Β Β // Debug UART
Β Β #ifdef DEBUG
Β Β common_hal_never_reset_pin(&pin_GPIO43);
Β Β common_hal_never_reset_pin(&pin_GPIO44);
Β Β #endif
}
My SCL and SDA are on GPIO41 and 42.
stuck elbow
are those somehow related to the debug uart? like, say, DTR or something?
(just making wild guesses here)
mental nexus
No idea, I didn't mess with this file.
stuck elbow
you have debug enabled?
mental nexus
Hmm.. Not sure. Unfortunately I'm just getting back to this after setting it aside a couple months ago. Let me review the latest version of the board.c file for this board and see if I there are hints. Thanks for the suggestion on where to look.
mental nexus
you have debug enabled?
Rebuilt without the DEBUG flag and works. π€·
Thanks @stuck elbow !
daring pumice
Ok. I think this means that CPy is working on my board without fault? Bear in mind that the 6 flashes are repeating over and over - that never stops.
daring pumice
My next question is whether or not I can include the code I've written in a .hex image and push it over to the board.
slender iron
Ok. I think this means that CPy is working on my board without fault? Bear in mi...
in a steady state it should blink every 5 seconds
continuous isn't working
it'd be much easier to debug with a debug uart at minimum
daring pumice
in a steady state it should blink every 5 seconds
Ok. Can I use any unused GPIOs for UART, or do they need to lead to specific pins on the chip?
slender iron
any will work
daring pumice
Ok, so I'd just declare them as UART in pins.c?
slender iron
CIRCUITPY_DEBUG_UART_TX in mpconfigboard.h
daring pumice
something like #define CIRCUITPY_DEBUG_UART_TX (&pin_Px_y) ?
daring pumice
Fair enough. Just to reiterate, I need this to work over Jlink - I don't have another interface here.
and it needs to be SWD
slender iron
that's not going to work
daring pumice
Can I do it over BLE?
slender iron
no. you need a UART for debugging at minimum. right now you are flying blind
daring pumice
Ok. I think my last option is to create a .hex image of my project with CPy and try to push it over to the board. I realize that's incredibly problematic, but it seems like the only path forward.
daring pumice
That's not exactly true - I'm able to step through the code line by line over GBD, and x-reference it with the code in the clone of the git repo.
daring pumice
it'll all be easier once you have usb
That won't be possible with the current revision of the board - there's a design error that we're correcting for the next revision, but I'm trying to save these prints for testing. :/
slender iron
trying to do it early is costing both of us extra time
daring pumice
I understand - I don't want to take all of your time here, and I'm grateful for the support you've given me so far.
slender iron
@tulip sleet or @onyx hinge did you pick up an s3 box lite?
daring pumice
I guess the best question I can ask at this point is (again) for guidance on how to create a .hex image with my code included. Is there a set of instructions I can follow posted somewhere?
slender iron
no, there isn't. you are well off the beaten path
daring pumice
Fair enough. Thanks again.
slender iron
circuitpython isn't designed to work as an appliance
tulip sleet
<@329766224093249548> or <@426200547024961539> did you pick up an s3 box lite?
not Lite; I got https://www.adafruit.com/product/5290 soem time ago.
slender iron
ya, I have that too. the lite is different enough to have a different board def
tulip sleet
should I get one? I though it was just without the dock
slender iron
nah. I can get one
I have a board def WIP that I was hoping one of you could finish
onyx hinge
I have a board def WIP that I was hoping one of you could finish
Same here I only got the original
daring pumice
Is it possible to prevent attempts to enable USB in CP? I think this is what's causing my board to continually reset.
slender iron
set CIRCUITPY_USB to 0
daring pumice
@slender iron you're a champ.
daring pumice
set CIRCUITPY_USB to 0
That works! I now get one blink ever 5 seconds or so!
timber mango
Hi everybody, I wonder if someone is interested in developped a library to play ogg vorbis or flac files on circuit python ?
manic glacierBOT
slender iron
@gilded cradle I'm around for two hours if you have time to debug the web workflow issue you are having
(I'm around off and on in the afternoon)
gilded cradle
<@478616919721771019> I'm around for two hours if you have time to debug the web...
Sure, we can meet up now. I was just about to get started on fixing esp-web-flasher, so I'd rather debug now instead of in the middle of that.
slender iron
can get you building ESP too
gilded cradle
I can't currently. I was going to work on spinning that up tomorrow
slender iron
I can walk you through it too
gilded cradle
ok, that'd be awesome
slender iron
@stuck elbow @tulip sleet want to chat here
tulip sleet
yah sure, so could we point to an empty sequence of some kind?
tulip sleet
but if not board.DISPLAY would do a truth-value test
so the idiom for testing would change
stuck elbow
using "is" correctly is really tricky
slender iron
mp_obj_is_true checks against == mp_const_none
if not board.DISPLAY is more flexible
tulip sleet
so instead of saying we set to None, we set to something else ... yes
slender iron
because we could add deinit to displays and then use that for true
setting it to something else means it prints that way
tulip sleet
so implement __bool__ for displays, then? I think
slender iron
that or for none
stuck elbow
either __bool__ or __nonzero__
I think the latter is the more pythonic one in this case
not sure if MP has it
stuck elbow
but it's not such hot code
slender iron
there is still a fast path for the None object
tulip sleet
that's a good point, it would fall through and still eventually be false
slender iron
right. it falls through now but the generic unary doesn't implement it
so copy the generic one and return false on bool
tulip sleet
it's still weird is None wouldn't work but it would print as None
slender iron
ya, I agree that's weird
stuck elbow
can we somehow raise an error if someone uses is with that fake None?
tulip sleet
if we added a deinit state, we'd need to check it in all the display calls at the beginning, as we do for busio objects, ec.
@onyx hinge ^
stuck elbow
another option is to have a NoDisplay object that is falsy
slender iron
can we somehow raise an error if someone uses is with that fake None?
I don't think so because it doesn't call out to the object
tulip sleet
I think NoDisplay or deinited are the cleanest
onyx hinge
I think it'd be interesting to have a 'generic deinitialized object', it might actually allow removal of a lot of deinitialized checks. but I once prototyped it and didn't save any code size, which was my goal then.
If we do something for displayio it would be nice if we tried to make it generic enough that we could use it for other cases too, like when you deinitialize a DigitalIO it turns into this type...
PumpkinType
stuck elbow
we could make it a generic NoDevice, then use it for other things too
slender iron
a generic deinit object would change the type out from under you
slender iron
for Displays yes
stuck elbow
you would then get errors like "A Display expected, but NoDevice found"
tulip sleet
NoDevice would respond to any method or attribute wit some exception, maybe RuntimeError, maybe OSError
maybe a NoDevice object has a single attribute that is the expected class
stuck elbow
raise OSError(418, "I am a teapot")
onyx hinge
Yeah I think you could make it say it's a <Deinitialized %q>. I should see if I still have that branch
onyx hinge
can't find it
stuck elbow
Yeah I think you could make it say it's a `<Deinitialized %q>`. I should see if ...
but then it can't be a singleton
onyx hinge
No, when you deinitialize any object that works this way, its memory gets overwritten so that it's a <Deinitialized %q>
stuck elbow
so you don't reclaim the memory?
tulip sleet
you can just copy the old type into the first field, maybe you did that
you could zero out the rest of the object, and then copy the type
or could alloc a new one
slender iron
you can't alloc a new one because you don't control pointers/references to the object
stuck elbow
so it has to be smaller than the smallest device object
slender iron
I think it is safe to assume an object is the type pointer plus at least two bytes
and each object type would have to deinit itself
so it would be opt-in so to speak
onyx hinge
stuck elbow
that's why we don't save on code size?
onyx hinge
deinited_unary_op should be changed so that use in a truthy context returns False
slender iron
lunchtime for me. the deinited object is interesting to me
onyx hinge
maybe I'll play with it for a bit while waiting on this other CI
tulip sleet
if someone could just summarize into the issue, that would be great; this is not an urgent fix, but is an interesting puzzle
onyx hinge
>>> i, bool(i)
(<I2C>, True)
>>> i.unlock()
>>> i.deinit()
>>> i, bool(i)
(<DeinitedType>, False)
>>> i.unlock()
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
ValueError: Object has been deinitialized and can no longer be used. Create a new object.
idle owl
@onyx hinge Hello. I am ready for the magic fix to resolve the destruction I unleashed on my CircuitPython fork. You mentioned it in passing, but not in a way that registered enough with me to recall it.
onyx hinge
<@426200547024961539> Hello. I am ready for the magic fix to resolve the destruc...
Ok I'm out at the moment so maybe in 30 minutes from now?
idle owl
Ok I'm out at the moment so maybe in 30 minutes from now?
I have a thing in 35 minutes. I'll ping you tomorrow if you're around.
main furnace
Can I disable the status line on CP8? RP2040 feather so I don't have WiFI.
jaunty juniper
you can hide the title group (though I think blinka and the status bar are not grouped together), and move the REPL group and reset the size of the console, but I think it's a little buggy right now, does something like that work ?
for pos in range(len(display.root_group) - 1):
display.root_group[pos].hidden = True
tile_height = display.root_group[-1].tile_height
delta = display.height % tile_height
supervisor.reset_terminal(display.width, display.height)
display.root_group[-1].x = 0
display.root_group[-1].y = - delta
Not yet. My intent is to add other status to it as well
so code running vs done vs error and USB status
main furnace
I'm using a 128x64 oled, I only have 4.5 lines of text.
jaunty juniper
Not yet. My intent is to add other status to it as well
speaking of which, can we put blinka and the status in a single displayio.Group ? that way it's always root_group[0] and the REPL is 1
slender iron
@jaunty juniper fine with me
@main furnace its meant to be useful π
but I understand it isn't now
main furnace
0.2 of the screen is a bit much, but for now I can keep that node on CP7.
slender iron
it may not be too much code to underlay the serial portion under the title
but then the title would be hard to read
@main furnace you can actually set the title from your code
so you could use that space yourself
(in theory, I haven't tried it yet)
main furnace
How does that work?
slender iron
it uses the OSC escape code
so try print("\x1b]0;my new title\x1b\\")
the \x1b is the escape key code
main furnace
Okay I'll give it a try. Thanks.
jaunty juniper
awwww no microcontroller.cpu.temperature on ESP32S3 ! So much example code dies in agony !
slender iron
idf support it yet?
jaunty juniper
I assume not ? I think we have updated the IDF since the first S3 support was added ?
tulip sleet
@jaunty juniper @slender iron it is in ESP-IDF master but not 4.4.x
I checked on this recently and re-checked. master has many changes from 4.4. I wouldn't try to go there now.
slender iron
π
tulip sleet
@onyx hinge I made little progress today on the ESP32 Thonny debugging. I think there may be some memory or heap error, but it's hard to tell. Adding logging creates Heisenbugs :/ . Tomorrow do you have time to get a brain dump about this, in case you want to work on it next week?
onyx hinge
@tulip sleet yes, before noon my time would be good
tulip sleet
got it, I'll ping you tomw morning, thanks
onyx hinge
OK!
jaunty juniper
<@199811382793863168> <@252717193496756235> it is in ESP-IDF master but not 4.4....
ah I see, thanks
orchid basinBOT
manic glacierBOT
manic glacierBOT
tulip sleet
@onyx hinge I am around for esp32 debrief pretty much any time today
onyx hinge
@tulip sleet OK, we're finishing entertaining guests soon and I'll ping you. maybe 30-45 minutes
onyx hinge
@tulip sleet OK guests are squared away until lunch time π
tulip sleet
ok, I am trying something; avail in a few minutes
onyx hinge
OK, I'll be around until 12 CT / 13 ET
@idle owl if/when you're around we can talk about undoing the changes in your fork of cirucitpython
tulip sleet
@onyx hinge ok, my experiment did not improve things, ready to talk. You can get latest from dhalbert/circuitpython, branch esp32
idle owl
@onyx hinge Are you still available to fix my fork?
It's ok if not. Like I said, it's not crucial yet.
onyx hinge
@idle owl not really, we are leaving as soon as Ingrid finishes up her call which was supposed to be 28 minutes ago π
idle owl
Ok, no worries.
onyx hinge
the general idea is locally you get the adafruit main branch and then force-push that to your fork's main branch though
idle owl
Hmm ok.
I'll try to figure that out. π Worst case, I guess I burn it to the ground and begin again.
onyx hinge
that's true!
thorny jay
I found a disturbing behaviour of the web workflow. If you set the .env, you get an IP but things are not working yet (except ping), but if you restart the board, then things start to work.
tulip sleet
I found a disturbing behaviour of the web workflow. If you set the `.env`, you g...
I noticed something similar. If you can specify the recipe for that, please file an issue. Scott is interested in all kind of behaviors, and expects bugs to be found
thorny jay
I noticed something similar. If you can specify the recipe for that, please file...
Boot a board, create a .env file, see the IP, try, fail, reset, try, it works!!!
jaunty juniper
or would we want the web workflow to only be started on reset ?
thorny jay
or would we want the web workflow to only be started on reset ?
Might be an option, don't know what is more/less confusing, but if instruction to reset are present, that's fine. I think having the IP but not the rest could be useful as it will be visible on the "console" and you are likely to have the same IP at next restart, so it help potentially. Anyway at one time you will have to set that .env file one way or another... or is there a trick for that too??? Like a configuration access point for initial setup?
About trick, plugging your board in a charger to have the file system Read/Write from wifi is a great one.
My workflow will include that, first plug in a power-supply, do your change, then plug it in the host when it will be used.
jaunty juniper
that could take the same form as the BLE workflow, press reset at a certain point during boot to get the board in access point mode, though maybe that ends up being a lot of boot modes
thorny jay
I must admit I skipped the BLE workflow, mostly because I don't have an iPhone handy, and if felt required.
thorny jay
that could take the same form as the BLE workflow, press reset at a certain poin...
NRF board are BLE only, S2 board are Wifi only... but then with S3 board you can have both and then the boot mode will become a bit complex. π¦
jaunty juniper
yup
wraith crow
@idle owl lol, thanks, was scratching my head big time
@thorny jay Can you try, boot board, create .env, see the IP, try, while waiting... create file and write to file ie t = open('test.txt','w'), t.write('test')
thorny jay
I need to give a bit of context... I have a piece of code take an "image" and encode it as mouse movement for a Nintendo Switch. Today I need to plug my CP board on a PC, copy the "image", then plug in the Switch, and I need a "button" to activate mouse movement (to avoid mouse movement while connected to the PC). My goal is to be able to only connect to the CP board with WIFI, add the "image" then restart (maybe auto-restart) and do everything without having to plug/unplug the CP board. So right now I plan to use and abuse from the Wifi Workflow so that I don't even need to implement a web server or anything like that.
wraith crow
Okay, I was just suggesting the file I/O because I wanted to see if an issue I ran into was causing your issue. When I recreate your steps on my board it does seem to apply. Basically any flash file I/O was unblocking the stalled web workflow. To confirm that with your board if you have host write access to the CIRCUITPY drive, all you need to do is rename a file on the CIRCUITPY drive. I don't mean this as a solution to your requirements, as the issue would eventually be resolved, just curious if it's the cause of your current observation.
thorny jay
Wait, in my problem I don't even have the Web interface working before doing the first reset after creating the .env file. π
Do you think writing to a file would wake it up???
wraith crow
But only for one operation at a time, so you may need to do multiple file i/o depending on what you're trying to do over web workflow
thorny jay
OSError: [Errno 30] Read-only filesystem
I think my problem is not what I initially described...
wraith crow
The file i/o can occur from either REPL or the Host PC via the CIRCUITPY drive
thorny jay
The Web Workflow fail for me on my QT Py S2 but is a success on the Feather TFT S3.
wraith crow
I haven't gotten around to seeing if my S2 boards see the same issue, but I first saw it on an S3 board
thorny jay
For me: S3 = no problem ... S2 = sometime ping, never the web interface.
wraith crow
I you're getting the Read-only error from a board that's plugged into a host and has a CIRCUITPY drive mounted, you could try creating a file or renaming a file from the host computer and see if that releases the web workflow
thorny jay
While I would live to have a minimal hardware for my usecase, I can develop everything on my S3 and figure out later what is going on with that S2.
wraith crow
This is all still very alpha, I'm sure it will end up working on both the S2 and S3
thorny jay
Nope, creating a new file, writing to an existing file. It does not like me.
wraith crow
You did the file access after initiating the webflow request?
thorny jay
I have a browser on one side and using touch or creating file with Mu. But I don't think the browser was requesting anything at that time.
I am trying to clean completely the board... but now the FS is read-only and I cannot delete a leftover secrets.json or edit the .env ... now I need a way to "format" my CIRCUITPY.
wraith crow
jaunty juniper
safe mode should always work
to be able to write from the web workflow, you only need to eject the drive, though I don't know if that works on windows
thorny jay
fsck to the rescue!
I was read only because of error on the filesystem. I tried the reset to factory default image, but it did not help. I failed at using esptool and the web serial also. Now I should be back to a clean QT Py S2. What a fight.
jaunty juniper
oh ok a corrupted drive
slender iron
ya, sorry for the bugs. I plan on going into bug hunt mode once I have websockets PRed
thorny jay
So the QT Py S2 is now clean, back with the latest version of CP. As soon as I set the .env it get an IP, but then whatever I do, I cannot get to the web interface. Only Ping is working for me. The good news is that I can play with the Web Workflow on my S3+TFT. So I know I can do it and I can try to experiment with the thing I had plan. Thank you for the help.
slender iron
I think you are right that you need to reset after writing .env
I'm not sure it checks after every auto-reload. I'll need to look at that
wraith crow
Doesn't the working ping mean that workflow has loaded?
slender iron
it means that its connected to the network but not that the webserver code is going
wraith crow
ah ok
thorny jay
The other good news is that with the bits and pieces of documentation I found and a few PR exchange, I was able to figure out what to do to use the Web Workflow, without having to watch a YouTube video.
thorny jay
Yeah it need to be retested from scratch once the other bug are solved. That will be the first contact for a lot of user.
slender iron
yup yup
I am assuming that most folks will set the .env when they load CP
especially for C3 and ESP32
thorny jay
I would have been great if the .env file and the secret.py would have more or less the same syntax and key-value. Maybe secret.py will not be needed anymore and everybody will use getenv, and a compatibilty secret.py can be created for having the credential in only one place. (not sure my idea are clear here but there is a way to get rid of secret.py I believe).
slender iron
ya, I'd like to replace secrets.py with .env
thorny jay
Unrelated: mdns or bonjour did not work for me, but it could be my OS despite installing avahi-daemon. It could also be my new wifi filtering multicast... too many variable for any serious troubleshooting... maybe I should start my Mac Mini. π
slender iron
you can test from the command line using dig
if you have two CP devices and can get the web workflow of one, the home page will try to do the mdns discovery for you too
wraith crow
New default secrets.py?
import os
secrets = {
'ssid' : os.getenv('CIRCUITPY_WIFI_SSID'),
'password' : os.getenv('CIRCUITPY_WIFI_PASSWORD'),
'aio_username' : "your-aio-username-here",
'aio_key' : 'your-aio-key-here',
'location' : 'New York, US'
}
slender iron
we should stop using secrets.py all together because it conflicts with a native CPython secrets module
wraith crow
π
thorny jay
if you have two CP devices and can get the web workflow of one, the home page wi...
both dig and nslookup fail to see circuitpython.local and I have now two board on the same wifi... that don't see each other. So I have to troubleshoot my wifi I am afraid.
slender iron
can you get the welcome page of one with the ip?
thorny jay
Yes I have both welcome page, but they show no other device.
slender iron
k, that must be the router then
thorny jay
Must be my super wifi blocking something. It is unify but brand new to me, and it is too simple for me (I was a cisco expert, that sexy web interface is disturbing to me). π
slender iron
I have a ubiquiti dream machine and its working ok
thorny jay
Some progress $ ping cpy-59d204.local PING cpy-59d204.local (192.168.192.72) 56(84) bytes of data. 64Β octets de 192.168.192.72 (192.168.192.72)Β : icmp_seq=1 ttl=255 temps=248Β ms
thorny jay
In Network Setting there is Multicast DNS and I had to enable that.
I may have changed the default value thinking that I was smarter than Ubiquiti. π
misty garnet
any particular reason there is no adafruit_bus_device for UART?
slender iron
usually uart is just one device
misty garnet
yeah, I was thinking that it would be useful for an enter/exit to open/close the port, but maybe that's a bit trivial
slender iron
I think you can use the UART object itself to auto-deinit
misty garnet
it looks like the UART exit will deinit, but that's not really what I was looking for
slender iron
what do you mean by close port?
misty garnet
hmm, good question, I could have sworn there was an open() and close() function but it appears I'm completely wrong
must be confusing it with the Serial class on the PC side of things
slender iron
Serial on iOS
spiral elk
So the QT Py S2 is now clean, back with the latest version of CP. As soon as I s...
I'm at the same state on ESP32-C3 (Lolin C3 Mini board). Have an IP that's pingable, mdns is working, but no one's home on port 80.
wraith crow
@spiral elk Are you compiling the Lolin C3 Mini build yourself or downloading a binary UF2 file?
wraith crow
If I put a build with a tweak on github would you be interested in testing it out?
spiral elk
My build env is outdated so it'd take me a bit to update it
wraith crow
I'll do the compile, if you're intereseted, let me know if a .bin will work and I'll post the URL
spiral elk
I'll do the compile, if you're intereseted, let me know if a .bin will work and ...
Heading to lunch now but I'll take a look at it once I'm done
C3 is bin only at the moment
And will probably always be so
No USB peripheral onboard
wraith crow
@spiral elk heh, I figured as much π
I'll probably have called it for the night so I'll check to see how you made out in the morning
spiral elk
I'll probably have called it for the night so I'll check to see how you made out...
That version worked first time
jaunty juniper
oh nice
wraith crow
@spiral elk The main difference in that version is that the code blocked mention in issue #6552 has been commented out.
https://github.com/adafruit/circuitpython/issues/6552
I really don't know if there are other impacts to Flash performance of running that version, but it's good to confirm the behavior. If you're interested you could update the issue with your results and I believe that will trigger notifications when progress is made.
languid whale
What is the touch display used in the 6049 PR movie?
proven garnet
This might have already been discussed, but if the intention is to use dotenv with .env files for storing secrets, is the plan for repo READMEs to keep suggestion making virtual environments with python -m venv .env or is the name of the environment going to change to something else? If I understand correctly, .env and .env/ wouldn't conflict, but didn't know if asking users to the "edit .env" would cause any confusion.
jaunty juniper
oh yeah they would conflict
you can do python -m venv .venv
I can't figure out what the bot doesn't like π€
I personally find .env a problem for beginners, the Finder won't let you create a dot file, it's impossible to save a file as ".env" with TextEdit (.env.txt at best)
ah lol
there's a trick to see it from the Finder (cmd-shift-G + ".env" in the directory) but otherwise it's command line (with the "open" command allowing to edit with TexEdit for example). I have not tested if Mu is ok with any of that
proven garnet
I'll bring it up during In The Weeds then
proven garnet
I don't have permissions for setting up webhooks via RTD, so could someone who does manage setting up the RTD project for this library: https://github.com/adafruit/Adafruit_CircuitPython_PCF8574
wraith crow
Is the workflow file browser supposed to show the SD card rather than the internal flash if an SD card is mounted?
wraith crow
If you mean, you wouldn't be able to modify the code.py file once an SD card was mounted, I'd have to agree. I would think there should at least be a way to toggle back to the internal, maybe there is and I'm just not seeing it yet.....
jaunty juniper
oh it should definitely not be showing the SD card
manic glacierBOT
slender iron
If you mean, you wouldn't be able to modify the code.py file once an SD card was...
Please file a bug. I suspect this has to do with which VFS we pick
jaunty juniper
Please file a bug. I suspect this has to do with which VFS we pick
we could do like os.listdir and such and pick the vfs based on the path
so we could list mounted sd cards
slender iron
we could do like os.listdir and such and pick the vfs based on the path
ya, let me think about it. handling read-only could be tricky
jaunty juniper
true
manic glacierBOT
lone sandalBOT
manic glacierBOT
manic glacierBOT
thorny jay
What do you think about the wifi banner that "print" on the REPL. In Mu it is a bit disturbing ]0;π Wi-Fi: 192.168.192.72\? What is the plan for this? Should Mu be updated to understand those and display them as a banner, or it should be disabled on the serial interface?
jaunty juniper
What do you think about the wifi banner that "print" on the REPL. In Mu it is a ...
there's a discussion here you can contribute to https://github.com/adafruit/circuitpython/issues/6548
since the title thing is a tty escape sequence I think it's not unthinkable to have Mu swallow it and possibly use it as status, but it might be good to disable it until that is settled
maybe have it be just normal text without the escape sequence in the mean time
or ctrl-something shortcut to toggle it ? would that make sense ?
thorny jay
Could be the "Prompt".
But if you have access and plan to use the REPL, either you don't need to know the IP or you have another way to get that info (I don't know witch, but there should be an API to know the IP currently in use, not sure to do what, but just in case).
That web workflow is still very confusing to me. π
jaunty juniper
yeah wifi.radio.ipv4_address does give the IP
digital shoreBOT
adafruit
manic glacierBOT
jaunty juniper
oh @idle owl I kind of need a sanity check here, I was trying to use the Sparkle animation and it doesn't work properly, the "sparkle" doesn't stay up for the length of a frame (as per the speed parameter), it's erased as fast as the code can, and certainly doesn't match the gif in the guide. When I look at the code I see the problem, but it looks like it's been like that for 2 years.
idle owl
oh <@330227457296957440> I kind of need a sanity check here, I was trying to use...
Huh. Weird. I guess the guide might have been written before the bug was introduced? Do you think you can fix the issue in the lib?
I can grab something and run it to check too before you dive in.
After meetings today though π
Because I've never had issues with it, but I also never updated the LED wall from my last office, so it was running older code the whole time. So I wouldn't have found bugs that way.
jaunty juniper
I can grab something and run it to check too before you dive in.
thanks, there's no emergency of course
idle owl
No worries! Good catch on your part if it's broken. I'd rather verify it and move forward with a fix sooner rather than later, if it's been identified.
onyx hinge
<@&356864093652516868> The weekly meeting is in about 1 hour. We love hearing what's up with y'all so it's a great time to update the notes document: https://docs.google.com/document/d/1tiCKK3V9-X9_jgHTmTlEjDPfB_z7bjZTNu2kYtM7KBw/edit?usp=sharing -- see/hear you soon
thorny jay
<@&356864093652516868> The weekly meeting is in about 1 hour. We love hearing wh...
Not making it to the meeting, maybe pre-read if my text is not too long or complex there are parts that can be easily skipped or summarised further.
idle owl
Not making it to the meeting, maybe pre-read if my text is not too long or compl...
I wouldn't be too concerned about it - we've all read off super long status updates from folks before. π
Thank you for the offer!
thorny jay
I'll make sure to listen to the recording. π
serene warren
Lurking. Trying to get back into the swing of things.
blissful pollen
Lurking too, my new job found out I exist (which is good) but harder to book lunch for the same time as the meeting
idle owl
It should be a Friday... looking now.
lone axle
idle owl
It is definitely a Friday.
If you want to contribute, please let me know, email circuitpythonday@adafruit.com !
inland tusk
It is a Friday for sure
lone axle
It was super cool. I could see a possibility in the future where the web editor / console could prompt you with an easy button to install missing libraries when it detects the import errors
jaunty juniper
It was super cool. I could see a possibility in the future where the web editor ...
that's a feature I was thinking of for Mu too. I've been wanting to integrate circup and a UI for circup in it, but making webapps in js is more of my strength than UIs in python
slender iron
I'd like to see use move code.circuitpython.org forward too
I'm hoping melissa is going to unlock it for us by making it work with more than ble
random junco
Pycharm and VSCode will detect .env and .venv
idle owl
I can manage patching that then to `.venv`
I'll get back to you on that π
slender iron
π good idea!
lone axle
I think restream.io has tools to help with streaming out a video that has been cut together as well
serene warren
Some people are introverts (yes me) and not comfortable with live stuff but might be ok with a short reording
random junco
It could be a demo or just a "Happy CircuitPython Day" message might be cool, too
idle owl
Some people are introverts (yes me) and not comfortable with live stuff but migh...
Exactly!
gilded cradle
Thanks
modern wing
Thanks for another great meeting!
[joined literally at the last minute due to other obligations -- but at least I kinda made it!]
serene warren
Thank you all for sharing your time
random junco
Thanks for another great meeting!
[joined literally at the last minute due to ...
You got a nice hug report.
mental nexus
Thanks everyone!
modern wing
You got a nice hug report.
Aww, that's nice to hear π
idle owl
Sprints! <a:super_love:407941462181937163>
Yeah, you're first on the shortlist to "host". π
modern wing
Here's towards a healthier tomorrow!
lone axle
Thanks for driving the ship today @onyx hinge. Great to have you back around and into the swing of things.
onyx hinge
Thanks!
minor plume
Iβm intrigued by the βsprintsβ idea for CircuitPython Day and would love to know more.
idle owl
Iβm intrigued by the βsprintsβ idea for CircuitPython Day and would love to know...
I'll keep you posted as it gets more clear what it might be π
minor plume
I'll keep you posted as it gets more clear what it might be π
Thanks! 8/19 is a Friday, and my company works half days on Fridays, so Iβm working on clearing my calendar for that day so I can be available after 1pm EST.
idle owl
Thanks! 8/19 is a Friday, and my company works half days on Fridays, so Iβm work...
Great!
blissful pollen
I'll have to listen to the end of the meeting but if some stuff is coming for CP day and requires help or such I can see how work is / book time off. Not hard to convince me not to work a Friday π
onyx hinge
Here is the notes document for next 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/1r0xVV98pjnzGh3ODbX7HxM_5_Tz8UaIbgFDG-5JZrAw/edit?usp=sharing
idle owl
@slender iron I can walk Tekktrik through our RTD process if you want.
slender iron
@idle owl thanks. I'm sure you know a more updated version than me
idle owl
At this point, almost certainly. π
daring pumice
Question re: Custom Board based on the Feather Sense:
The board I've printed isn't supplying the right amount of power to the VBUS (it's giving 3.3V when it should be giving 5V).
I have a variable external power supply that I've used to compensate (bring the board's voltage up to the right level).
CPy boots at that point with no issue, but it still doesn't show up as a drive that I can access (i.e. can't see it as a drive, can't push my code over to it).
daring pumice
I'm sure I'm missing something. Does anyone have any ideas or guidance on how I could save this board?
slender iron
Generally a USB sniffer is useful when determining if anything is happening over USB.
lone axle
I noticed a difference in the way that these attributes render when built locally with sphinx vs. what is showing on the live RTD page: https://docs.circuitpython.org/projects/logging/en/latest/api.html
As far as I can tell I am using the same version of sphinx 5.0.2 as what RTD built with. Anyone know of what else could account for the differences? Ultimately I was hoping to make the name and type have a space (or something else) between them. But was puzzled when I found that my local build doesn't look the same.
slender iron
maybe theme version?
idle owl
As far as I can tell I am using the same version of sphinx `5.0.2` as what RTD b...
Uffffffff. I've run into this as well. And it always turns out to be something obtuse that takes ages to find.
lone axle
checking on sphinx-rtd-theme now. I do think I had a different version
idle owl
Like I had indented lines not working right, and they only worked right with 1 space indent. 2 or more made them show up completely wrong. Took ages to figure that out, because Sphinx explicitly states, any number of spaces works fine.
lone axle
yep, sphinx-rtd-theme was the culprit. 0.4.3 is the version that I found an RTD build used. I was on 1.0 I'm at least matching the live one now so I can figure out how to tweak it.
And a whitespace tweak (remove a space) is also the fix to make it not get smashed together as well in this case. Thanks for point in the right direction to both of you.
manic glacierBOT
slender iron
@wraith crow what board are you seeing this issue on?
wraith crow
I'm pretty sure it happened on an UM Feather S3, but I've been running through my S3 and S2 boards so it might have been an S2. I've noticed it at least twice but it's not very repeatable.
slender iron
kk. I suspect it is worse on C3 since it doesn't have USB to keep things moving
wraith crow
I'll try and test out a C3 later tonight then π
slender iron
just made the websocket PR βοΈ
wraith crow
The times it has happened I renamed or deleted a file from the flash and the workflow recovered, it was definitely hung for more than 5 seconds though
wraith crow
No, the workflow isn't responding when it happens, either from the CIRCUITPY drive or from the serial REPL
slender iron
hrm, ok. so you change it over USB
wraith crow
yep
It looks like my FeatherS2 is in the hung state now, is there anything you want me to try?
slender iron
are you running a debug build?
are you using a web browser to access it?
viewing the network tab of the developer tools in the browser can help see what request is in flight
wraith crow
No, I don't think it's a debug build, yes I'm trying to access it from a browser
slender iron
did anything load or are you waiting for the html?
it'd be helpful if you could get debug output as it happens
wraith crow
slender iron
the debug info prints out what requests its gotten
do you have it open in other tabs?
I wonder if it is chrome
tries chrome
I did have issues where an old socket would block new ones but I thought I fixed that
wraith crow
I hooked up rx/tx and it looks like I am on a debug build, but the output gets mucked up after the cpu starts....
slender iron
hrm. maybe different speed?
I'm on S3 mostly
just saw weirdness chrome so I'm pulling that thread
wraith crow
Looks like a baud rate issue, but I've never seen it change in the middle like that
I'm going to delete a file from flash and see if it kicks the device loose
Yep, from the REPL I deleted a file from flash (boot_out.txt) and now the web workflow is responding
slender iron
huh! weird
wraith crow
The symptoms are exactly the same as what was happening when the code block I commented out was still in play, except that when the code block was there, performing flash I/O only allowed a single webflow transaction and you had to keep performing I/O to get anything done. Now when it locks up, one or two I/O unlock it and then it runs without issue.
slender iron
ah, the hook that runs the workflow is tick dependent
I didn't realize that from the function name
I haven't been doing USB transactions so I haven't really hit this
I wonder why ticks are on at the start
aka, why does it work for me
wraith crow
I was thinking it had something to do with the filesystem_dirty flag. I'm running on an old Windows 8 machine and I'm not always dismounting the CIRCUITPY drive cleanly
But that was a very uninformed guess π
slender iron
I think by commenting out the tick disable, you are leaving it on
and it calls the "background" task that does the http serving
but I'm not sure why it is on for me at all
wraith crow
I wasn't sure how it was working for anyone either, something else must turn the ticks on that wasn't in play in my setup
slender iron
I have an idea how to set it up to be driven by the sockets
but it isn't super simple unfortunately
ah! I'm on a board with a display
and displayio turns on ticks
wraith crow
Ah ha π
slender iron
turns off the display
wraith crow
Foamyguy likes display boardss as well, I was scratching my head watching his streams....
jaunty juniper
I tried that excuse with my boss once, it didn't go well
wraith crow
Sleep is good!
wraith crow
Hopefully I'll have a chance to play with your new PR tonight π
slender iron
do you want a patch so you don't need to do USB stuff?
wraith crow
I can throw my patch in after I get your PR
slender iron
somewhere in init you need an extra supervisor_enable_tick
wraith crow
It's working well enough for me to play tonight, it can wait π
slender iron
kk np
midnight ember
Oh I got a mention for my bitmapsaver example in the meeting today. My first ever PR. Thanks for the mention. β€οΈ Sorry for using the SD card @lone axle didn't mean for you dig out an SD card just to confirm it. I could have written it using SPI flash instead. It's cooler with the SD card example though, can log endless screenshots with SD. π
lone axle
Oh I got a mention for my bitmapsaver example in the meeting today. My first eve...
No worries. It was in a v1 of a 3D printed case with no slot for the SDCard access. I need to print a newer one before I get it closed back up.
midnight ember
oh no, well, made you want to print a new case. one of those circumstances where's it fine until it's not fine. i'm working on a 3D printed version with someone for a tft featherwing case. secret project. no SD card access though, just thought of that. :/
just seeing the PR comments now, yeah i completely assumed the 2.4" was drop in compatible. good call on the different driver. :/
tiny token
Not sure how to move forward with this, so I'll pose the question here. I've continued to tinker with the HTTPServer by @tulip sleet. In my fork I have two branches (Refactor & WSGI). The first one refactors the module in to a folder with class files, the second builds on that by adding support for the CGI and WSGI spec.
jaunty juniper
sounds good so far
tiny token
My goals. 1. allow the example code file to run unchanged 2. Keep HTTPServer simple so beginners can use it to quickly standup a web server on their boards. 3. Allow adventurous developers to use the WSGI pieces to create more advanced web apps.
Still a ways to go, in getting it cleaned up (docs, comments, etc.) but I've got it working fairly well so far. I've tested it on an ESP32S2 and examples work unchanged for both branches.
so besides the Docs, and comments. What else should I be looking at to make sure that these changes are acceptable? Is updating HTTPServer the best way to do this or should it be a separate module for the community bundle?
blissful pollen
My goals. 1. allow the example code file to run unchanged 2. Keep HTTPServer ...
Two thoughts. One if you are free Monday's 2pm EST attend the weekly CP meeting held on discord and bring it up in the "in the weeds" section, where longer discussions on items just like that take place.
Two, you could create a draft PR or an issue (tagged with enhancement) and explain what you are trying to do and create discussion there.
Three (yes I know I said two) I think Dan is away right now but a lot of times here on Discord as a good place for discussion as well.
wraith crow
@slender iron Heading to bed but wanted to say it was really cool seeing the REPL come over the Websocket. This is really exciting stuff, Thanks!!!
tiny token
Two thoughts. One if you are free Monday's 2pm EST attend the weekly CP meeting ...
Two (3) great ideas. :-) I think I'll start with the issue/enhancement that'll give me a place to organize the thoughts. Mondays are tough because of work. I'll add it to my calendar though and try to make some of them.
Need to dig into this web socket stuff too, similar, I believe, just used for different purposes but maybe useful. Sigh!... So much to learn so little time.
manic glacierBOT
manic glacierBOT
onyx hinge
@slender iron not really! I had one idea but I looked and it was already done π π
there may be optional parts but we don't like turning off optional parts
afaik we configure for 2D which is the common case, but you might make sure of that
higher dimensions bring in more code
slender iron
my websocket build failed because the ulab update added 2k+. I'll look at what my options are
onyx hinge
yes it certainly got bigger
slender iron
only two boards are too small though
dan and I looked a bit when you were out and saw ulab is pretty big
should I make GIFIO depend on camera?
slender iron
@wraith crow you may want to use a more sophisticated terminal emulator too. mine is very basic
wraith crow
what terminal emulator is that?
slender iron
it is my custom code. basically a <pre> tag π
only filters out the terminal title, backspace and clear line codes
wraith crow
I'm not sure I'm seeing exactly the performance you meant since I swapped the javascript method but it seems to work pretty well. Command recall (up arrow) would be nice but I can't really think of too much else. I'm seeing an occasional "Null" inserted, I think it's when I overwhelm the buffer or something, but it hasn't bothered me too much yet.
slender iron
I think up arrow requires me to track keypresses
you can do beforeinput on chrome. I just had to change how to connect it
(testing it now)
wraith crow
So on an older PC I'm picking up the network name of the device from my router (umfeathers2), I can ping that name and bring up the web workflow welcome page. When I go to one of the pages that needs authentication (umfeathers2/fs/ or umfeathers2/cp/serial/), the page isn't displayed and the DEBUG console comes up with a CP webserver: bad origin http://umfeathers2 message
slender iron
ah, that is because you used the DHCP hostname rather than the MDNS one
wraith crow
I don't think I have MDNS on Windows 8 :/
wraith crow
Yep, the IP works fine
slender iron
there is an origin_ok function in web_workflow that does the check
wraith crow
What happens if you have multiple of the same type boards, does MDNS/DHCP enumerate them showhow?
slender iron
trying to prevent rogue http sites on the web from accessing it
MDNS should
I'm not sure about dhcp because the hostnames may be identical
the mdns ones use the mac address to be unique
wraith crow
I've seen DHCP devices with MAC addresses stuck on them, so maybe if it detects a collision it adds it
slender iron
Β―_(γ)_/Β―
onyx hinge
(inlined by) mp_obj_print_helper at /home/jepler/src/circuitpython/ports/espressif/../../py/obj.c:133
(inlined by) mp_obj_print_helper at /home/jepler/src/circuitpython/ports/espressif/../../py/obj.c:114
fun_builtin_1_call at /home/jepler/src/circuitpython/ports/espressif/../../py/objfun.c:75
mp_call_function_n_kw at /home/jepler/src/circuitpython/ports/espressif/../../py/runtime.c:665
mp_execute_bytecode at /home/jepler/src/circuitpython/ports/espressif/../../py/vm.c:936
fun_bc_call at /home/jepler/src/circuitpython/ports/espressif/../../py/objfun.c:297 (discriminator 4)
mp_call_function_n_kw at /home/jepler/src/circuitpython/ports/espressif/../../py/runtime.c:665
mp_call_function_0 at /home/jepler/src/circuitpython/ports/espressif/../../py/runtime.c:638
parse_compile_execute at /home/jepler/src/circuitpython/ports/espressif/../../shared/runtime/pyexec.c:146
pyexec_friendly_repl at /home/jepler/src/circuitpython/ports/espressif/../../shared/runtime/pyexec.c:734
run_repl at /home/jepler/src/circuitpython/ports/espressif/../../main.c:823
(inlined by) main at /home/jepler/src/circuitpython/ports/espressif/../../main.c:922
app_main at /home/jepler/src/circuitpython/ports/espressif/supervisor/port.c:410
main_task at /home/jepler/src/circuitpython/ports/espressif/build-adafruit_feather_esp32_v2/esp-idf/../../esp-idf/components/freertos/port/port_common.c:141
``` ooh you can get even more information of addr2line if you ask fori t@slender iron do you think anyone is depending directly on the output format of ./tools/decode_backtrace.py ?
onyx hinge
sadly, gdb disassembles mp_obj_get_type incorrectly. The faulting instruction is at +35. ``` 0x400dec50 <+28>: j 0x400dec5a <mp_obj_print_helper+38>
0x400dec53 <+31>: movltz a3, a0, a0
0x400dec56 <+34>: lsi f2, a5, 0x39c
0x400dec59 <+37>: .byte 0xff
0x400dec5a <+38>: l32i a8, a10, 16
did you try and print out the pointer passed into get_type?
wraith crow
Is there an standard way to do case insensitive string compares when writing CircuitPy C code?
slender iron
not that I know of
wraith crow
kk π
manic glacierBOT
bleak matrix
Hi. There is a build failure for a single board on the main branch. Although it appears unrelated to my change, is it possible for me to re-run this GitHub action?
import adabot.github_requests as github
ModuleNotFoundError: No module named 'adabot.github_requests'
Error: Process completed with exit code 1.
digital shoreBOT
adafruit
jaunty juniper
In some Magtag learn guides, like this one:
https://learn.adafruit.com/magtag-weather/project-code
β adafruit_minimqtt is missing from the project bundle
β it is required for the adafruit_io library
β the requirement is not spelled using the pypi formatting
β that should not be an issue for pypi (it converts _ to - and ignores case)
Maybe that makes the project bundler miss that requirement ?
Actually that seems to affect every guide using adafruit_io (so any using portalbase)
Like that:
https://learn.adafruit.com/pyportal-weather-station/code-pyportal-with-circuitpython
manic glacierBOT
manic glacierBOT
manic glacierBOT
I think I found the problem with psram. Now, thonny works with the adafruit feather esp32 v2!
slender iron
Hi. There is a build failure for a single board on the `main` branch. Although i...
Generally we just comment it is unrelated and merge anyway
will look shortly
bleak matrix
Thanks!
slender iron
@onyx hinge I'll look shortly
onyx hinge
@slender iron a re-review would be awesome
proven garnet
@idle owl Did you ever write the guide about mpy files? It'd be really helpful for my understanding while reviewing a PR, but I can't seem to find it.
idle owl
<@330227457296957440> Did you ever write the guide about mpy files? It'd be rea...
I thought so. Hold on.
Thank
idle owl
Oh it's in the welcome to CPY guide, excellent! I'll remember that. π
FWIW, it took me a minute to remember that.
jaunty juniper
@idle owl I noticed adafruit minimqtt missing from the project bundles on learn for the portalbase libraries
I believe this PR will fix it https://github.com/adafruit/Adafruit_CircuitPython_AdafruitIO/pull/93
but I don't have access to the bundler, so I don't know
idle owl
@jaunty juniper That PR will fix it, if that's the issue you were seeing.
@proven garnet Any reason not to merge that PR?
You're welcome to π
jaunty juniper
ok that's good then
proven garnet
You're welcome to π
ADD, best reason I got. I'll merge it now π
slender iron
@onyx hinge I'm waiting for a re-review from you too
onyx hinge
@slender iron oops, will do that now!
slender iron
thanks!
onyx hinge
That's certainly an odd failure. python3: can't open file '/home/runner/work/circuitpython/circuitpython/ports/atmel-samd/../../tools/uf2/utils/uf2conv.py': [Errno 2] No such file or directory
idle owl
ADD, best reason I got. I'll merge it now π
Hah! Fair enough.
onyx hinge
seems the "get CP deps" step failed but didn't report the error in a way that GA understood as a failure
proven garnet
Also, @idle owl whenever you get a chance, could you take a look at this cookiecutter fix? It's for building MPY files for community libraries, which is currently broken (sometimes, PR should explain) https://github.com/adafruit/cookiecutter-adafruit-circuitpython/pull/186
Oh cool it has merge conflicts, I'll fix those
idle owl
Oh cool it has merge conflicts, I'll fix those
Merge when ready.
proven garnet
You got it!
manic glacierBOT
orchid basinBOT
proven garnet
@idle owl, @tulip sleet, @jaunty juniper - I was thinking about how to add __version__ info to the libraries, and it fits really well with pyproject.toml. I have an example in my CircuitPython_CSV library that I just converted to pyproject.toml. It uses the same sed command to populate the version as a hardcoded string instead of using setuptools-scm. That way, any changes to the library don't make it deviate from the tag name upon release (PyPI absolutely hates tag names like 1.2.4+dev.1234556 and won't do it) but this side steps the problem. Then in a future update, we could add __version__ to all the libraries that don't already have them and you could do this across the board:
>>> import libname
>>> libname.__version__
1.2.3
PR with the changes for reference: https://github.com/tekktrik/CircuitPython_CSV/pull/13
And it works, the release I just did for v1.1.0 worked: https://pypi.org/project/circuitpython-csv/
slender iron
@strong violet please ping me when you are around. I want to modify thonny to work with the set title stuff in CP 8 and would love some pointers
slender iron
@lone axle do you have simple editor html/js?
lone axle
<@382939733107408897> do you have simple editor html/js?
yes
slender iron
do you want to integrate it into the web workflow or want me to?
slender iron
fine with me. you'll likely hit build size issue unfortunately
we'll have to decide how to make it fit
slender iron
@onyx hinge I did the resolve
lone axle
The branch is here https://github.com/FoamyGuy/circuitpython/tree/webworkflow_edit_page I'm heading out for a walk, but can work on a PR for it tonight afterward.
slender iron
kk, I'll do other things π
onyx hinge
<@426200547024961539> I did the resolve
thank you!
onyx hinge
@slender iron maybe we can look at it together tomorrow, I enabled web workflow on esp32 v2 and it does get an IP address (I can see it in the repl), I can ping circuitpython.local but but I can't browse or curl http://circuitpython.local
slender iron
do you have a display?
otherwise you'll need 6591
(which I'm cleaning up now)
don't know how I missed all that debug stuff
onyx hinge
like displayio? no
slender iron
as-is in main you'll need ticks enabled
onyx hinge
I see, I think
slender iron
right now the call to the http server code is done from a ticks callback
which works with displays because displays enable ticks
onyx hinge
that branch will have a minor merge conflict with esp32 PR as well
slender iron
ok, I've got fixes for it now anyway
maybe I'll wait until the ESP32 PR is green and then merge
57 checks left
manic glacierBOT
onyx hinge
nice!
onyx hinge
I don't see any files once I get /fs to load. when there's no password configured, /fs and /serial both just give blank pages, not guidance. /serial does work in ff now that I configured the http password
>>> os.listdir('/')
['.fseventsd', '.metadata_never_index', '.Trashes', 'lib', 'boot_out.txt', 'adafruit_ahtx20.py', '.env', 'wifitest.py']
```files are seen in the replthis is really slick @slender iron well done!
manic glacierBOT
slender iron
@onyx hinge my intent was that fs/ wouldn't load at all without a password
slender iron
1 more esp32 check...
jaunty juniper
oh @idle owl a note about learn guides: while responding to support questions, I noticed that there are a few featherwing guides that have list of pins like this:
https://learn.adafruit.com/adafruit-adalogger-featherwing/pinouts
The SDCS pin is the chip select line.
On ESP8266, the SD CS pin is on GPIO 15
On ESP32 it's GPIO 33
On WICED it's GPIO PB5
On the nRF52832 it's GPIO 11
On Atmel M0, M4, 328p or 32u4 it's on GPIO 10
On Teensy 3.x it's on GPIO 10
They should probably be updated for new boards like RP2040, ESP32Sx feathers, TFT and such (most of which use the same pinout as the M4, so maybe put that upfront) and mention the CP pin name (like board.D10 in this example) or something
idle owl
oh <@330227457296957440> a note about learn guides: while responding to support ...
Can you please email me with this info? I'm out until Monday and this will get lost on Discord.
jaunty juniper
will do
onyx hinge
<@426200547024961539> my intent was that fs/ wouldn't load at all without a pass...
yes I agree, the password is a great idea .. but I'm suggesting that when there is no password configured that it needs to do something besides return a blank page, like give a hint to configure the access password in .env
slender iron
Ya, we can tweak it all. Just takes more firmware space. π
onyx hinge
yep
pulsar bloom
cb@speedy:~/circuitpython/ports/espressif$ make BOARD=adafruit_feather_esp32s2_tft
Use make V=1, make V=2 or set BUILD_VERBOSE similarly in your environment to increase build verbosity.
Traceback (most recent call last):
File "../../tools/gen_web_workflow_static.py", line 7, in <module>
import minify_html
ModuleNotFoundError: No module named 'minify_html'
Trying to build the latest
manic glacierBOT
jaunty juniper
cb@speedy:~/circuitpython/ports/espressif$ make BOARD=adafruit_feather_esp32s2_t...
have you pip install -r requirements-dev.txt ? It's a new requirement
manic glacierBOT
pulsar bloom
Thanks, builds now!
pulsar bloom
cool repl in one tab file browser in anotherπ
manic glacierBOT
warm stump
https://github.com/TheMindVirus/whistles
TODO(The): Rename Whistles.eval(const char* src); to Whistles.interpret(char* repl);
Bells and...undefined references
midnight ember
Was told by tannewt to post here for help in adding a flash chip definition for the nrf52840 feather bluefruit sense. https://github.com/adafruit/circuitpython/tree/9d5b04ebc49193afab8efba4fb31c33ac1c81f5d/ports/nrf/boards/feather_bluefruit_sense I'm going to attempt to drop in a bigger QSPI flash chip (16MB). https://www.digikey.com/en/products/detail/cypress-semiconductor-corp/S25FL128LAGMFM010/7322845 I checked the datasheet and far as I can tell it appears to be drop in compatible. It hasn't arrived from digikey yet and I'm pretty excited to get started before it arrives. I can do a PR but unsure where to start. Should I fork the main Adafruit/circuitpython or something deeper like /ports/nrf/boards? Where should my fork start? My experience with contributing on GIT is almost non-existent.
idle wharf
@midnight ember have you been through the building CP learn guide?
I donβt consider myself a developer and with the excellent docs from the CP team Iβve been able to make contributions, like adding new boards.
I am sure you can do it. Ask questions as you go. Community here is great!
midnight ember
first time i'm seeing that, probably because i don't do much with ESP... yet (I have 2 S3's on the way). thank you. π
idle wharf
Thereβs another learn guide on adding boards as well. Youβre instinct is correct, copy a similar board and start with that.
midnight ember
documentation looks good, i've never worked with toml before. i forsee pip installs in my future.
should i name the exact model in the filename? https://github.com/adafruit/nvm.toml/tree/main/flash/cypress
current ones are much shorter S25FL216K.toml vs the one i'm getting S25FL128LAGMFM010
EXTERNAL_FLASH_DEVICES = "S25FL116K, S25FL216K, S25FL128LAGMFM010, GD25Q16C" no way, is it really going to be as easy as putting a comma and chip id? that's too easy, there's gotta be some catch.
midnight ember
created a fork of adafruit/circuitpython and it's showing me this message
Your main branch isn't protected
Protect this branch from force pushing, deletion, or require status checks before merging.
should i click on protect now or just ignore it?
i have no other contributors in my personal repo so don't think i have to worry about it, just double checking.
spiral elk
`EXTERNAL_FLASH_DEVICES = "S25FL116K, S25FL216K, S25FL128LAGMFM010, GD25Q16C"` n...
Adding chip to a board is indeed that easy if its in nvm.toml
midnight ember
it's not there, i added that part to the line
spiral elk
Then you need to stare at datasheets and make new entries most likely
midnight ember
yep, prepared for that.. tomorrow. it's late.
or early, i haven't been to sleep yet. lots of research to do π
spiral elk
Often similar parts are fine, GD25Q16C also worked for me for GD25Q32 parts
i.e. the chinese pico clones with 4MB of flash.
manic glacierBOT
Define the default BLE name of a board, for BLE workflow in particular (for example in a classroom setting).
This is only the default name, user code can later redefine the BLE name too.
Use CIRCUITPY_BLE_NAME in .env.
CIRCUITPY_BLE_NAME=Bill's Circuit Playground
radiant trench
@slender iron I was just watching Show and Tell. I've been looking for a way to get more involved in CircuitPython but I'm not much of a Python developer. That said I do have a lot of experience with CSS. Scott, would you at all be interested in letting me collaborate on the web workflow stuff you're doing by providing some super minimal CSS that you could include to jazz up the interface a little?
radiant trench
(btw by "Jazz it up" I mean, make it more readable and usable than the browser defaults) π
manic glacierBOT
slender iron
(btw by "Jazz it up" I mean, make it more readable and usable than the browser d...
I'd love help! The challenge is doing it without increasing the file sizes much. They are here: https://github.com/adafruit/circuitpython/tree/main/supervisor/shared/web_workflow/static Most should be workable from your desktop and communicating with a local device
radiant trench
I'd love help! The challenge is doing it without increasing the file sizes much....
Yeah i imagine space is pretty limited. Iβm pretty sure I can make it a little nicer without too much overhead. Especially once the css is all minified.
So I guess I can just do a pull request or something with my changes?
Donβt suppose it would be acceptable to assume the user has internet and pull the css from a cdn?
slender iron
my goal is to not assume internet however if it works ok without I think we could have a supplement css for online
ya, PR is great and I can help you get testing
right now the files are gzipped. html and js is minified using some python libraries
I'll be back in 20 or so. gotta change a diaper
slender iron
@wraith crow you around?
wraith crow
Sort of, taking a break by the pool at the moment
slender iron
kk, what board are you testing on? I can make you a build with more debug prints
and you can try it when you have time
wraith crow
I've been using the unexpected maker feather us too but I also have an adafuit ESP 32s2 and an S3 dev kit
slender iron
kk
wraith crow
Um feather s2, π voice transcription didn't get that very.well
onyx hinge
aha thonny can "show hidden files", I just didn't look in the correct place at first. There's a hamburger menu in the "open from circuitpython device" dialog, and that's where you can do it... and it seems to be remembered, at least in the session. yay, now I can edit .env easily.
jaunty juniper
how would I debug print an int ? I've been using serial_write like a savage
slender iron
how would I debug print an int ? I've been using serial_write like a savage
mp_printf(&mp_plat_print, βformat stringβ, input, variables);
onyx hinge
What is a 'station'? that is, what does start/stop station method on wifi.radio do? Is there a way to undo wifi.radio.connect without soft-resetting?
jaunty juniper
I wonder what they do too, I believe a station is the term for the "client" side that connects to an access point, Micropython has network.STA_IF and network.AP_IF
sta_if = network.WLAN(network.STA_IF)
if not sta_if.isconnected():
sta_if.connect('<your SSID>', '<your key>')
I know of switching wifi.radio.enabled to force a disconnect I think ?
wraith crow
@slender iron So the behavior starts out the same, no DEBUG messages, nothing on Dev Tools Console, Network tab shows pending connection to 10.0.0.187
If I connect via the USB serial port, the Auto-reload message is waiting for me with a message to press any key to enter the REPL.
Pressing enter causes the following on the DEBUG console:
I (186319) CP socket: opened 57
I (186319) CP webserver: Request GET /
I (186319) CP webserver: Header Host 10.0.0.187
I (186319) CP webserver: Header Connection keep-alive
I (186319) CP webserver: Header DNT 1
I (186329) CP webserver: Header Upgrade-Insecure-Requests 1
I (186339) CP webserver: Header User-Agent Mozilla/5.0 (Windows NT 6.3; Win64; x64) AppleWebKit/537.36 (KH
I (186349) CP webserver: Header Accept text/html,application/xhtml+xml,application/xml;q=0.9,image/avi
I (186359) CP webserver: Header Accept-Encoding gzip, deflate
I (186359) CP webserver: Header Accept-Language en-US,en;q=0.9
The browser displays the welcome screen with no fields filled in, Still nothing on the dev tools console, the network tab shows:
slender iron
The two pending are sockets waiting to be accepted
wraith crow
Pressing Ctrl-D on the repl at that point kicks out the following DEBUG messages:
I (544469) CP socket: opened 58
I (544469) CP socket: closed 57
I (544469) CP webserver: Request GET /favicon.ico
I (544469) CP webserver: Header Host 10.0.0.187
I (544469) CP webserver: Header Connection keep-alive
I (544479) CP webserver: Header User-Agent Mozilla/5.0 (Windows NT 6.3; Win64; x64) AppleWebKit/537.36 (KH
I (544489) CP webserver: Header DNT 1
I (544499) CP webserver: Header Accept image/avif,image/webp,image/apng,image/svg+xml,image/,/*;q=0.
I (544499) CP webserver: Header Referer http://10.0.0.187/
I (544509) CP webserver: Header Accept-Encoding gzip, deflate
I (544519) CP webserver: Header Accept-Language en-US,en;q=0.9
Still nothing on the dev tools console, the network tab shows the favicon.ico download complete, still pending on welcome.js
midnight ember
using a similar part won't work in my case because there's no 16MB chips listed for cypress in the nrf52/bluefruit sense. i don't think there are any boards for nrf52 with that sized chip. i know it exists for the teensy and some espressif boards. i can use another cypress chip as an example and resize the flash parameter hopefully.
wraith crow
Pressing Enter at the new REPL press any key prompt results in:
I (718779) CP socket: opened 57
I (718779) CP socket: closed 58
I (718779) CP webserver: Request GET /welcome.js
I (718779) CP webserver: Header Host 10.0.0.187
I (718779) CP webserver: Header Connection keep-alive
I (718789) CP webserver: Header User-Agent Mozilla/5.0 (Windows NT 6.3; Win64; x64) AppleWebKit/537.36 (KH
I (718799) CP webserver: Header DNT 1
I (718799) CP webserver: Header Accept /
I (718809) CP webserver: Header Referer http://10.0.0.187/
I (718819) CP webserver: Header Accept-Encoding gzip, deflate
I (718819) CP webserver: Header Accept-Language en-US,en;q=0.9
and:
onyx hinge
Possibly a red herring but out of ESP32, -S2 and -S3 isn't S2 the one with just one core?
slender iron
@onyx hinge c3 is single too iirc. I was going that route too. S2 failed quickly in mdns code for me. So Iβll try and take a look at that once Iβm back to my desk
onyx hinge
ah ok
slender iron
Ari has an eye doc appointment now so itβs a busy day
tidal kiln
@onyx hinge fwiw. decent cheat sheet.
slender iron
is back
slender iron
love it when debug prints fix an issue
onyx hinge
huh I just noticed the IP address shows as blank ..
version.json has ip: ""
this probably happens if the terminal title is disabled
manic glacierBOT
midnight ember
i don't see a way to easily append a bigger chip to mpconfigboard.h #define CIRCUITPY_INTERNAL_NVM_SIZE (4096) if i change that value then i'd have to create a new board, was hoping to avoid that and just get a bigger flash chip supported. am i going to have to make my own board to add a bigger flash chip and if so what should i name it?
this is for the bluefruit sense https://github.com/adafruit/circuitpython/tree/main/ports/nrf/boards/feather_bluefruit_sense
hmm found an example from samd that uses if statements for different variants. https://github.com/adafruit/circuitpython/blob/main/ports/atmel-samd/mpconfigport.h says it's set within include/sam.h could i do something like that?
jaunty juniper
hmmm the nvm size would be the size of the nvm, not the external flash (microcontroller.nvm)
midnight ember
oh oops. ok then this one #define BOARD_FLASH_SIZE (FLASH_SIZE - 0x4000 - CIRCUITPY_INTERNAL_NVM_SIZE)? trying to change the size of the spi flash. didn't realize nvm was different than external flash.
i'm guessing nvm is internal to the nrf52840 chip?
midnight ember
# Settings for the Cypress (was Spansion) S25FL128L 16MB SPI flash.
# Datasheet: https://www.infineon.com/dgdl/Infineon-S25FL256L_S25FL128L_256-MB_(32-MB)_128-MB_(16-MB)_3.0_V_FL-L_FLASH_MEMORY-DataSheet-v09_00-EN.pdf?fileId=8ac78c8c7d0d8da4017d0ed40e335224
total_size = 0x200000 # 2 MiB
start_up_time_us = 10000
memory_type = 0x40
capacity = 0x15
max_clock_speed_mhz = 133
6b_quad_read = true
32_qspi_write = true``` I've found and changed the max clock speed but the rest is a mystery to me. I don't understand the MiB conversion. Can't find anything mentioning startup time, no clue what the memory or capacity should be in hex.
dusk mauve
I haven't run an espressif port build in a few months on my dev box... Is there anything I need to do other than a git pull? Like does the esp-idf in the ports dir need to be updated or anything?
blissful pollen
you make need a make fetch-submodule (or similar I may have that wrong) as I think the IDF was updated at somepoint. And then to run the install.sh script again. I think that's what I had to do
dusk mauve
you make need a `make fetch-submodule` (or similar I may have that wrong) as I t...
Thanks, that's exactly what I needed to hear π