circuitpython-dev

[manic glacier]

dhalbert

[circuitpython:2\.x] 2 new commits

6b66c3b Add W25Q16 USON flash support, alternative flas... - ladyada
544b9e4 Merge pull request #541 from ladyada/itsybitsy_... - dhalbert

[manic glacier]

ladyada

[adafruit/circuitpython] New comment on pull request #541: Itsybitsy m0 fixes

uhh yeah, thats totally on purpose and not because i accidentally added support to master by accident and then had to backtrack!

[jovial wind]

we had a few instances of a CPX somehow getting its CircuitPython scripts wiped while demoing today...I suspect it was due to a bit of restarting the power to calibrate the tab capacitances

[stuck elbow]

@jovial wind like this? https://github.com/adafruit/circuitpython/issues/386

[jovial wind]

@stuck elbow yep, exactly like that

oh, it's a low power thing? This was with a lipo I've had lying around just for testing things and not bothered charging, then used it a fair bit today...so was likely low on charge when it started happening

[stuck elbow]

It probably happens when you are connecting the power, but it's not immediately good, because the connection is flaky or whatever.

[solar whale]

@bronze geyser FYI - I was able to get the RFM69 library working with an RFM69 featherewing on a feather_m0_adalogger (non-express) the only change to simpletest.py was to set the RESET and CS ```# Define pins connected to the chip.
CS = digitalio.DigitalInOut(board.D10)
RESET = digitalio.DigitalInOut(board.D11)

[tulip sleet]

@jovial wind we added some code to guard against this power problem in CircuitPython 2.2.0. Upgrade to that if you haven't already. If it's already 2.2.0, then we may need to do more work to prevent this kind of thing, so let us know

[upbeat plover]

I've had 2.2.0 crash on me more then once where it gives a Whoops error msg, and I have to put the .uf2 back on the CPX

[pulsar bloom]

I am working on getting the Parallax 360 servo working and making headway. Getting good data for 180 degrees then it is off in the weeds.

[solar whale]

@upbeat plover - just a followup to a previous conversation - I have been able to to load both the RMF69 dirver Si7021 sensor driver on a feather_mo_express. I can get a reading from the si7021 and send it out via the RMF69 featherwing.

@upbeat plover what were you doing when it crashed?

[upbeat plover]

trying to get it to load code with lines with code over 80 char long

once i started using MU to check that it stopped

it would give an error randomly somewhere where it shouldnt, then after a long wait it would load the whole code sometimes... was horrible.... checking your code in MU is nice

[pulsar bloom]

Code for test Parallax servo, do not understand pulses.resume() with a parameter.

[solar whale]

@upbeat plover ow were you loading the code - before mu.

[upbeat plover]

ive been using atom for a while, it doesnt tell you stuff like that

just recently i downloaded MU... everytime I launch MU it gives an error... kinda annoying but worth the built in REPL and code checker

[solar whale]

recently starting using mu - it is nice.

[upbeat plover]

"Could not find an attached Adafruit CircuitPython device." I thought CPX doesnt need a driver for windows 10?

the REPL works fine after launch so it obviously is finding it lol

[solar whale]

I don't have a windows system - I get that on my linux box if I have something else connect to the devixe - like screen. keep forgetting to exit it first. ONly one connection allowed at at ime!

are you starting mu before powering or connecting the device? It has to be there first or you will get that error.

[upbeat plover]

gives error both ways

[solar whale]

hmmm - odd but it sounds like it it working OK

[upbeat plover]

Yeah, its working fine, just a little annoying hearing the windows error sound at launch and the error msg. Maybe I should report a bug?

[solar whale]

sure - maybe someone else can comment on if they also see that.

[upbeat plover]

knowing windows it probably is only effecting me 😢

hmm found a work around, before closing MU if i change it from adafruit to python on the bottom right corner, when I launch next it wont give the error, but i have to switch it back to adafruit before clicking REPL or MU crashes

[stuck elbow]

affecting

[opaque patrol]

@upbeat plover what version of Mu are you using?

[upbeat plover]

1.0.0.beta.13

i got a different version

now

nvm that was for bbc microbit

[opaque patrol]

There was an issue with Mu beta 12 on Windows with the COM ports that was fixed in beta 13 (thanks @meager fog )

[upbeat plover]

i think i may have found the issue

yup, CIRCUITPY wasn't the drive name

[pulsar bloom]

eureka! probllem solved, had to buffer more readings even though I was only using the first two to give Circuitpython time to pause before the buffer was over run.

[raven canopy]

import pulseio is failing with no module found. Feather M0 Express, CP3.0.0-alpha1.62
n/m. just saw the issue for ASF4...back to 2.2! 😄

[timber mango]

@solar whale

 $ microcom -s 115200 -p /dev/ttyACM0

I have very good results using microcom to connect to the m0 target boards. Plays nice with Arduino IDE. Can still upload to the target without disconnecting from microcom first (it just gets overridden somehow). Some other terminal programs also exhibit this nice behavior, but not all of them.

microcom may be a Rob Landley creation. https://landley.net

[stuck elbow]

the "nice behavior" is broken locking

[timber mango]

Is that a bad thing in this simple instance? @stuck elbow

[stuck elbow]

wait, you are talking about the m0

that can upload independently of the serial console

only esp8266 has that problem

[timber mango]

 $ $ ldd /usr/bin/microcom 
        linux-vdso.so.1 (0x00007ffd91045000)
        libreadline.so.6 => /lib/x86_64-linux-gnu/libreadline.so.6 (0x00007fd4e935d000)
        libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007fd4e9140000)
        libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007fd4e8d95000)
        libtinfo.so.5 => /lib/x86_64-linux-gnu/libtinfo.so.5 (0x00007fd4e8b6b000)
        /lib64/ld-linux-x86-64.so.2 (0x00007fd4e95a7000)

[stuck elbow]

did you really have a problem uploading to the m0 and using the serial connection at the same time?

because it should just work, with any terminal program

[timber mango]

The problem was that the serial program did not terminate and had to be manually exited and restarted, iirc.

[stuck elbow]

why would you need it to terminate?

[timber mango]

Because it would be a dead connection and no longer responds.

[stuck elbow]

interesting, I'm sure that would be a bug

and I think it's pretty much fixed in the recent versions of cp

you can have problems with the esp8266 boards (HUZZAH), though, because it uses the serial for uploading

[timber mango]

This applies mainly to Arduino. Sorry. ;) Let me ask you (that's telling to ask not asking to ask, hehe) ..

Is the baud rate parameter ignored when passed on the command line to start a session with CPX, Feather M0 Express &c? Seems to be irrelevant.

[stuck elbow]

it's a USB-emulated serial connection, it doesn't use the baud rate

[timber mango]

What is the equivalent baud rate of such a connection? What kind of throughput does one expect?

[stuck elbow]

not sure

I'm not even sure where to look for this

[timber mango]

Well for example CP2104 establishes a device in Linux by virtue of plugging it in. There's a kernel module for that.

[stuck elbow]

yes, but cp2104 has a real serial connection on the other side

which does use the baud rate

in case of the samd21 on the m0 boards, the usb connection is directly to the microcontroller

[timber mango]

Right. The Linux kernel must make sense of whatever is electrically present on its USB port. I have no idea how it does this. ;)

[stuck elbow]

there is a serial device usb spec

[upbeat plover]

what characters are supported for MU? some ASCII characters cause MU to crash

[stuck elbow]

@upbeat plover sounds like a bug

[timber mango]

Here's dmesg on this:
http://pastebin.ca/3961689 (expires in 2 hours)

cdc_acm I don't know about but it's obviously associated with /dev/ttyACM0

[stuck elbow]

cdc is a device class for the usb

[timber mango]

 $ lsmod | egrep cdc
cdc_acm                30362  0 
usbcore               195468  11 snd_usb_audio,uvcvideo,ums_realtek,usb_storage,cp210x,usbserial,snd_usbmidi_lib,ehci_hcd,ehci_pci,usbhid,cdc_acm

[upbeat plover]

it is a horrible bug, when i went to save MU crashed and the file i was working on got shrank to a zero sized file.... i have to start over no back up on it

[stuck elbow]

acm is "abstract control model", I guess it's a "catch all" class?

@upbeat plover I would report it if I were you

@timber mango this might be helpful: https://rfc1149.net/blog/2013/03/05/what-is-the-difference-between-devttyusbx-and-devttyacmx/

[timber mango]

GTKTerm exhibits the unwanted behavior vs microcom. The lockfile is there, but the terminal itself is unresponsive after an Arduino upload.

[stuck elbow]

I don't know about Arduino

I'm only using it with CircuitPython

and I know that CP doesn't reset the USB connection when you upload

[timber mango]

 $ Found upload port: /dev/ttyACM0
/home/nis/.arduino15/packages/arduino/tools/bossac/1.7.0/bossac -i -d --port=ttyACM0 -U true -i -e -w -v /tmp/arduino_build_530005/Infrared_Read.ino.bin -R 

That's what's klobbering GTKTerm. I think. Or whatever follows afterward. There is probably a series of steps one can take within GTKTerm to recover, without exiting, but I find microcom much simpler as it just goes away after this upload.

[stuck elbow]

well, yes, arduino pretty much replaces all firmware on the board, it has to reset everything

[timber mango]

I will see what I can generate using CircuitPython on this same hardware, but tbh I wasn't thinking about CircuitPython when I nudged jerryn. ;)

[stuck elbow]

sorry, I just assumed CP because that's what I use

[timber mango]

It is my error entirely. You are always very good about things and it does not go unnoticed by me. ;)

[solar whale]

I'm happy with screen and nano. Occiaonlly use mu. No problms copying files to M0 boards when screen session is open. My point was taht there can only be one REPL session open - Either screen (or equivalent) or mu.

[idle owl]

Yep, I've run into that too. Sometimes have to reset the board if I close screen without actually exiting it.

[timber mango]

screen is fairly specifically designed to be able to establish multiple shells from a single connection to the host. Ctrl A C (create) and Ctrl A N (next) iirc are the two main keybindings. There's a detatch keybinding; you can leave screen running on the far-away system (server) detached. People used it to maintain a presence on a telnet-reachable BBS while not actually online themselves (having detached the screen session and then closed the telnet (or ssh) session out to the ISP's shell server).
I don't think I've ever used it as a serial comm program! Never crossed my mind it could do that. ;)

[idle owl]

That's pretty much all I use 😃

[timber mango]

It's a nice terminal emulation and is worthy without using its virtual terminal features. Been around forever, and very unixy -- Sun Microsystems SunOS (later, Solaris) running on their SPARC server machines had a nice implementation, for example.

[timber mango]

Ctrl_A d will detatch screen (in Linux).

 $ screen -RR

to resume.
Also, to recover from a CTRL_A z sequence:

 $ fg

does the job.

To find a detached screen (very unusually) you can't 'ps | egrep screen' -- has to be SCREEN (I remember this; it's tradition).

 $ ps auxwwwww | egrep SCREEN  # << **this succeeds**
nis  12605  0.0  0.0  26072  2972 ?      Ss 20:54 0:00 SCREEN /dev/ttyACM0 115200
nis  12924  0.0  0.0  12732  2156 pts/16 S+ 21:06 0:00 grep -E SCREEN
 $ ps auxwwwww | egrep screen   # << **this fails**
nis 12899  0.0  0.0  12728  2080 pts/16  S+  21:04  0:00 grep -E screen
 $ pgrep screen # << **this succeeds**
12605

[solar whale]

When I run this code on an esp8266 ```import adafruit_sdcard
import busio
import digitalio
import board
import storage

Connect to the card and mount the filesystem.

spi = busio.SPI(board.SCK, board.MOSI, board.MISO)
cs = digitalio.DigitalInOut(board.GPIO2)
sdcard = adafruit_sdcard.SDCard(spi, cs)
vfs = storage.VfsFat(sdcard)
storage.mount(vfs, "/sd")

Use the filesystem as normal.

with open("/sd/test.txt", "r") as f:
print(f.readline())

I get this errorAdafruit CircuitPython 2.2.0-5-g544b9e4 on 2018-01-20; ESP module with ESP8266

import sdread
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "sdread.py", line 12, in <module>
File "adafruit_sdcard.py", line 360, in readblocks
File "adafruit_sdcard.py", line 238, in _block_cmd
File "adafruit_sdcard.py", line 207, in _cmd
RuntimeError: maximum recursion depth exceeded

[idle owl]

I've gotten that when I had a loop in the wrong place, or an indent got borked.

Is the code changing between files somehow?

From pasting it or something, did you end up with code in the wrong place. Guessing based on what I've done.

[solar whale]

hmmm - I'll double check. It works if I do it manually but withput the context manager fot the open.

strange - this works ```import adafruit_sdcard
import busio
import digitalio
import board
import storage

Connect to the card and mount the filesystem.

spi = busio.SPI(board.SCK, board.MOSI, board.MISO)
cs = digitalio.DigitalInOut(board.GPIO2)
sdcard = adafruit_sdcard.SDCard(spi, cs)
vfs = storage.VfsFat(sdcard)
storage.mount(vfs, "/sd")

Use the filesystem as normal.

#with open("/sd/test.txt", "r") as f:

print(f.readline())

f=open("/sd/test.txt", "r")
print(f.readline())
f.close

[idle owl]

Huh

[stuck elbow]

The recursion limit for mp/cp is quite low

Around 10 I think

[solar whale]

Is it different for esp8266 from m0? where is the problematic recursion in the posted code?

[stuck elbow]

It's called recursion limit, but it actually applies to any function calls

You can't go deeper than the limit

[solar whale]

OK - I'll see if I can find a difference in the builds.

[stuck elbow]

But from the traceback it seems there are only 5 calls

[solar whale]

Right ....

[bronze geyser]

@tidal kiln ...about mpy-cross-2.2.0-macos-high-sierra -> I am running Sierra, not High Sierra . Also, I assume I'd run this in a terminal window. However this file does not have an extension? My goal is to be able to take a .py made on Sierra and convert to .mpy. (?)

[stuck elbow]

Executable files don't usually have extensions on unix systems

[idle owl]

@bronze geyser mpy-cross doesn't have a file extension, it's a compiled program run from commandline. So you place it in a directory, and then run it on the .py file you're tyring to convert.

[solar whale]

@stuck elbow do you recall where the recursion limit is set?

[idle owl]

@bronze geyser So it'll look something likempy-cross yourfile.py

[bronze geyser]

@idle owl hi non-intimidating person. Here's the challenge... I downloaded mpy-cross-2.2.0-macos-high-sierra ... i assume this runs on Mac OS w/in Terminal. I open terminal in directory where mpy-cross-(really long name) is. I do mpy-cross-(really long name) adafruit_rfm69.py ... i then get command not found....

[idle owl]

Was the download a zip? (let me look)

[timber mango]

 $ ./mpy-cross_foo

the dotslash helps it find it in the current context. ;)

[solar whale]

@bronze geyser try two things chmod +x mpy-cross.... ./mpy-cross.... yourfile.py

firs line makes it executable -- but first unzip if its a zip file!

[timber mango]

 $ file ./mpy-cross

This tells you what this file is.

[idle owl]

It's not a zip, I checked

[stuck elbow]

@solar whale no, sorry, but I would look i mpconfig.h or portconfig.h

[bronze geyser]

./mpy-cross: Mach-O 64-bit executable x86_64

@timber mango thank you for the tip.

[solar whale]

@stuck elbow no problem -- I'm digging - tahtnks

[stuck elbow]

@solar whale or just grep for RECURSION_LIMIT

[idle owl]

@bronze geyser Yep it's definitely not executable when it's downloaded. So follow what @solar whale said. chmod +x mpy-crossYou can rename the longer file to mpy-cross to make all of this easier if you like.

[bronze geyser]

hmm... i guess High Sierra exes don't run on Sierra (and yah - i renamed the file, thank you).

[solar whale]

@stuck elbow intereseting - found an entry for samd ```jerryneedell@Ubuntu-Macmini:~/projects/adafruit_github/circuitpython_2x$ grep -r "RECURSION_"
atmel-samd/asf/sam0/utils/preprocessor/mrecursion.h:#ifndef MRECURSION_H
atmel-samd/asf/sam0/utils/preprocessor/mrecursion.h:#define MRECURSION_H
atmel-samd/asf/sam0/utils/preprocessor/mrecursion.h:#define MRECURSION_LIMIT 256
atmel-samd/asf/sam0/utils/preprocessor/mrecursion.h: * range from 0 to MRECURSION_LIMIT.
atmel-samd/asf/sam0/utils/preprocessor/mrecursion.h:#endif /* MRECURSION_H */

[stuck elbow]

That may be unrelated

[idle owl]

@bronze geyser I did the chmod +x and then used ./mpy-cross dir/to/myfile.py and it worked, however, I am running HS. I have another machine running Sierra, so if no one else gets to it first, I can build you an appropriate mpy-cross, but I won't be able to get to it until this evening.

[timber mango]

 $ ldd ./mpy-cross
# (produces a lot of output here)

Should confirm if an (otherwise) executable file makes sense to the operating system or not. Not sure what it outputs when you have a file that cannot be run properly.

[solar whale]

ah - but taht is in the asf code, not really helpful

[stuck elbow]

try grepping for the exception

You will find the code that throws it

Sorry I'm sick in bed with a tablet

[idle owl]

@stuck elbow Aw! I hope you feel better soon!

[stuck elbow]

Thanks

[bronze geyser]

i assume you meant ld? here 's what I got 'ld: warning: -arch not specified
ld: warning: -macosx_version_min not specified, assuming 10.11
ld: can't link with a main executable file './mpy-cross' for inferred architecture x86_64'

[idle owl]

I think you're right that it won't work with Sierra.

Some of them were backwards compatible, but this may not have been one of them.

[solar whale]

Binary file build/py/stackctrl.o matches
Binary file build/py/qstr.o matches
Binary file build/firmware.elf-0x09000.bin matches
Binary file build/firmware.elf matches
Binary file build/firmware-combined.bin matches
build/genhdr/qstrdefs.generated.h:QDEF(MP_QSTR_maximum_space_recursion_space_depth_space_exceeded, (const byte*)"\x73\x1e\x20" "maximum recursion depth exceeded")
build/genhdr/qstr.i.last:    mp_raise_RuntimeError("maximum recursion depth exceeded");
build/genhdr/qstrdefs.preprocessed.h:Q(maximum recursion depth exceeded)
jerryneedell@Ubuntu-Macmini:~/projects/adafruit_github/circuitpython_2x/esp8266$ 
```   hmmm

[slender iron]

@solar whale it coudl also be a function of the stack space in the port

[bronze geyser]

@idle owl i guess i should move to HS. I am concerned to do so since I run on 2013 macbook and do not want to upgrade hw (I really like it).

[timber mango]

@bronze geyser ld is related, but ldd is distinct (and correct here).

[bronze geyser]

@timber mango that's interesting I get ldd: command not found

[stuck elbow]

@slender iron so a function with a lot of local variables?

[idle owl]

@bronze geyser I finally had to upgrade my 2012. Don't upgrade to HS just for this. I can get you the right file.

[timber mango]

Not necessarily a bad thing, HD. I'm on Linux so not all commands are exactly one-for-one with Mac stuff.

[stuck elbow]

Do closures also live on the stack?

[slender iron]

@stuck elbow most python objects end up on the heap I think

[bronze geyser]

@idle owl i just didn;t want to be a pain. If you can provide file, greatly appreciate it. In no hurry...i have three projects going and switch between as i get stuck...thank you.

[solar whale]

@slender iron I was afraid of that - so is it an issue or "feature" - just learn to work around it?

[slender iron]

file an issue, it might be an easy fix for 3.x

[idle owl]

@bronze geyser You're not a pain! Promise. It's not a problem at all.

[slender iron]

we haven't looked much at the memory layout of the esp

[stuck elbow]

It's a bit crazy

[timber mango]

Since it's an mpy can't someone compile an .mpy for HappyDay and just post it here?

[bronze geyser]

@idle owl - as always thank you for your kindness. Please let me know when I can grab the file.

[slender iron]

yeah, I remember reading the post about esp memory mapping etc

[idle owl]

@bronze geyser Will do!

[stuck elbow]

On kickstarter

That was a long time ago

[slender iron]

ya

[solar whale]

OK - I'll create an issue. and stop scratching my head... thanks @slender iron @stuck elbow

[slender iron]

I think I'm onto something with this heap stuff. this latest run had a free run of 2288 bytes out of 6k free. It used to be 400 😃

thanks @solar whale

[stuck elbow]

Woot

[slender iron]

will post a new video once its done, rendering the frames now

[solar whale]

@bronze geyser BTW - there is an mpy for the rfm69 driver https://github.com/adafruit/Adafruit_CircuitPython_RFM69/releases

It's in a .zip file

[idle owl]

@bronze geyser See if this works 😃

[solar whale]

@idle owl much better for @bronze geyser to have a working mpy-cross ...

[idle owl]

@solar whale Yeah I had one less than a month ago on my Sierra machine from a CP build, so hopefully it works. I figured I'd need to do another build, but that's pretty recent.

[solar whale]

should be fine. I only think it changed with the major releases.

[idle owl]

Oh right... bugger. No that was before 2.x so it might return incompatible.

@bronze geyser You might get an error when you try to use the .mpy that creates saying "incompatible .mpy, please update". If you do, let me know, and when I'm back in a few hours I'll get you a new version. It won't be a problem, just let me know!

You won't get the error until you try to run the .mpy file, you'll need to check the serial output to find it. If it ends up being there.

Be back later!

[timber mango]

(had an idea)

[stuck elbow]

Ideas are dangerous

[manic glacier]

jerryneedell

[adafruit/circuitpython] Issue opened: #542 Recursion error on ESP8266

WHn I run this code on an ESP8266 it fails as shown below. It runs successfully on an M0 (feather_m0_adalogger)

import adafruit_sdcard
import busio
import digitalio
import board
import storage

# Connect to the card and mount the filesystem.
spi = busio.SPI(board.SCK, board.MOSI, board.MISO)
cs = digitalio.DigitalInOut(board.GPIO2)
sdcard = adafruit_sdcard.SDCard(spi, cs)
vfs = storage.VfsFat(sdcard)
storage.mount(vfs, "/sd")

# Use the filesystem as normal.
with ...

[bronze geyser]

@idle owl WoWza! (sorry for delay - my daughter came home and i was checking out the color she died her hair)... the thing i didn't get was i needed to ./mpy-cross (vs just mpy-cross) at the command line. A HUGE thank you.

[timber mango]

That's because 'dot' isn't in your $PATH:

 $ echo $PATH

The dot means current working directory (and dot dot means the one above it).

Apparently, I don't use dot in my $PATH anymore; I assume it's not in my path and always use the ./file syntax (which just basically specfies the current directory as the path to file).

Also, if I use an executable file (a command) a lot, I put it in my personal bin directory, which is in my $PATH (in fact my personal bin directory is the first thing in my $PATH).

[idle owl]

@bronze geyser That's great!

[bronze geyser]

i am having a challenge getting the adafruit_rfm69.py library working on a Feather M0 RFM69. Here: https://learn.adafruit.com/adafruit-feather-m0-radio-with-rfm69-packet-radio/pinouts#rfm-slash-semtech-radio-module it notes CS = 8. But in my python code, dir(board) notes there is no D8 ... so I can't set CS = digitalio.DigitalInOut(board.D8) ..'object has no attribute D8' ?

[idle owl]

@bronze geyser What does dir(board)give you?

[raven canopy]

@bronze geyser I don't have a RFM69, so i can't verify, but I imagine if you follow @idle owl 's direction you'll find an entry like RFM69.CS. It is defined in the pin settings of the RFM69 firmware. https://github.com/adafruit/circuitpython/blob/4bd1d09d74ba341796829145fd165c357fbeec4f/ports/atmel-samd/boards/feather_m0_rfm69/pins.c#L28

[jovial wind]

@tulip sleet ok, will do

[slender iron]

here is the latest heap video: https://www.youtube.com/watch?v=5uBF-bDXs8g

its getting close, just one or two more things to long-live

[idle owl]

Wow

[sharp bramble]

Does anyone know where I should look to find documentation for using the accelerometer on the CPx? My google-fu is failing on this one.

[sharp bramble]

Or, I have the library on the board (I can see the LIS3DH library), and I've attempted to use examples from the source (https://github.com/adafruit/Adafruit_CircuitPython_LIS3DH/blob/master/examples/accel.py) and this... simplified lesson (https://learn.adafruit.com/circuit-playground-lesson-number-0/accelerometer), neither of which yield anything. I am certain I'm missing something obvious.

[idle owl]

Are you using the Circuit Playground or the Circuit Playground Express?

[sharp bramble]

Express.

[tulip sleet]

@sharp bramble in CircuitPython? Use this library: https://github.com/adafruit/Adafruit_CircuitPython_LIS3DH or use convenience functions in https://github.com/adafruit/Adafruit_CircuitPython_CircuitPlayground

whoops I was too slow typing

[idle owl]

What Dan said. The link you posted is specific to Circuit Playground, so the code won't be the same.

[sharp bramble]

Yes.

There's a lot of truth in that statement.

[idle owl]

Here is a project using the CPX and the cpx class version of acceleration: https://learn.adafruit.com/ufo-circuit-playground-express?view=all#code-the-ufo-with-circuitpython

[sharp bramble]

from adafruit_circuitplayground.express import cpx

It's unclear where in the overview docs that line exists.

[idle owl]

You need the adafruit_circuitplayground/express library in your /lib folder on your CPX

[sharp bramble]

Essentially, I want to wrap some convenience functions to provide a functional Makecode -> Python experience for my students. I don't know where I was supposed to find the documentation that would have introduced that class/library?

I have it in the lib folder; the setup is correct. My challenge is knowing what M to RTF, essentially.

http://circuitpython.readthedocs.io/projects/circuitplayground/en/latest/adafruit_circuitplayground/index.html?highlight=express

That.

[idle owl]

Ah then you want cpx. The most thorough set of info we have on that class is https://circuitpython.readthedocs.io/projects/circuitplayground/en/latest/adafruit_circuitplayground/index.html#module-adafruit_circuitplayground.express

[sharp bramble]

That was the page I couldn't find.

Or, as you just linked. Thank you. I really do appreciate it.

[idle owl]

RTD isn't super beginner friendly, but that page has examples for everything available in that class

[sharp bramble]

I was pounding through the readthedocs pages for accelerometer/accel/etc., and could not find anything that pointed me at higher level documentation that works.

The bummer here is that I'm a college prof who has done embedded work in the past. I'm going to have to write docs for my students.

However... I shall license it freely!

I'll probably find my way back here again, but in the meantime, thank you, kattni.

[idle owl]

For sure! Always feel free to ask questions here!

[sharp bramble]

IT JUST WORKS

Now, I'm going to dig into that library.

[idle owl]

Excellent!

[solar whale]

@bronze geyser the CS Pin on the feather M0 RFM69 is not available to you directly as you have noticed. @timber lion has been working on getting a special version of Circuit Pyton released just to deal with this. IN the meanwhile, there is a workaround - I'll post some instructions as soon as I can do some testing of them.

[manic glacier]

invisibleroads

[adafruit/circuitpython] New star added

[solar whale]

@bronze geyser better yet, here is a build for the RFM69

@bronze geyser if you flash this to your board do as @idle owl suggested and execute import board dir(board) you should see RFM69_CS, RFM69_RST. THeses are the RESET and CS pins to use in the rfm69 test code.

@bronze geyser there is another workaround you can try instead of flasjing the ne w version if you prefer.

@bronze geyser try adding import microcontroller then use# Define pins connected to the chip. CS = digitalio.DigitalInOut(microcontroller.pin.PA06) RESET = digitalio.DigitalInOut(microcontroller.pin.PA08)

This should work with your existing version - it uses the "internal" pin definions .

[stuck elbow]

would be nice if that build also included the library as built-in

I mean, as frozen

[solar whale]

@stuck elbow That would make it easier to use, but would also then need bus_device, correct?

[stuck elbow]

I guess so

[solar whale]

@stuck elbow I guess it's a trade off about where to put the impact of the library. It takes up a significant portion of the files system. If it is "frozen" it will make the image bigger. I have to go back and look at the build - Is the File system a fixed size or is it just what is left after the image is created?

[stuck elbow]

It's fixed

Has to be because of how fat is organized

[solar whale]

OK so the "frozen" modules just fill "wasted" space - if they can fit.

[stuck elbow]

they probably can't, unless we remove something

It's all tradeoffs

[solar whale]

yeah - there was ony 4Kbyte free in the flash for that builld.

[stuck elbow]

The compuled library is about 10kB

[solar whale]

oops

[stuck elbow]

Without the bus_device

[solar whale]

this is a real limitation of the non-expess boards...

[stuck elbow]

Yes, that's why the express ones were created :)

[solar whale]

you can add an sdcard, but then there is probably not enough RAM for the radio driver and the sd driver 😦

[stuck elbow]

A driver written in C could be much smaller, and could use interrupts

[jovial wind]

interrups! 😃

[raven canopy]

vector check: for a rotary encoder, which approach would be better: upper-level pulseio library written in circuitpython, or at core-level as part of pulseio (e.g. pulseio.Encoder())? I'm thinking SAMD21 only for right now, but I know that SAMD51 has built-in (PECD) so am including future incorporation in my design thoughts.

[languid sage]

@idle owl coolAPI doc... do yo think in the future you might be able to put what rev of the software it refers to, or at least a "last revised" date. In a moving target environment like CP, this might be handy.

Thx

[bronze geyser]

@solar whale Thanks. Works w/ feather_m0_rfm69_firmware.bin and @idle owl - the mpy-cross to squeeze the wasted space out of adafruit_rfm69.py ...although I do admire how well the rfm69 file is commented. Out of the total number of 752 lines, 305 (roughly 43%) are comments. (umm..yah, well i'm learning regular expressions...).

[solar whale]

@bronze geyser the rfm69.bin should be included in future releases. It has been merged into the repository, but has not been released. Glad it is working for you!

[bronze geyser]

@solar whale I am very grateful that you took time to give me access. but then...perhaps I am greedy because I already have my mind set on LoRa....

[solar whale]

No problem! It's great to have someone test it! I dont have the feather_m0_rfm69 - I just have the featherwing versions so its uses a differnt pin for CS.

[idle owl]

@languid sage At the bottom of the sidebar, there's "Read the Docs" and a green button that says "v: something" with a little arrow, that allows you to swtich between latest and stable - which gives you an idea, but is vague. The way to check specifically at the moment is check the rev found at the bottom "Revision 9679cd8c" and then view it on GitHub and check the release information: https://github.com/adafruit/Adafruit_CircuitPython_CircuitPlayground/releases (which is the same as the revision). With CircuitPython iteself, it's more obvious from the "v" button at the bottom, because it lists the different branches, and is therefore not vague at all. But with libraries, we don't have different versions. For the most part they're backwards compatible unless otherwise noted in a release. So if you're concerned about compatibility, I would suggest checking the release page on GitHub for more information. RTD is still evolving for us, so it's possible we'll end up working in a revision date at a later time.

[sharp bramble]

From CircuitPython on a CPx, am I able to set up an interrupt on buttons A and B? Specifically, I'm interested in recreating "was_pressed()"-type behavior from Makecode. Or, would I need to be at a lower level on the tower... as in, that support would need to be in the runtime, or be implemented in C?

I might be able to fake it with a fast timer interrupt checking button state, but that's polling, again...

[idle owl]

Actual interrupt support isn't available in CircuitPython yet. So whatever route you choose will have to be something similar to an interrupt. I'm not entirely sure what that would look like. I've worked with interrupts only peripherally so I don't have a lot of information on how to deal with them.

[sharp bramble]

Good to know.

My first solution was to write watch_buttons(), and have students put that at the top of a while True loop.

[idle owl]

It's likely something that will be added eventually, but it's not high on the priorority list for core work at the moment.

Ok yeah that makes sense

[sharp bramble]

They can then call was_pressed(BUTTONA), and it will return true if the button was pressed. It buries some debouncing.

I'll keep my eye on the repo for developments.

Thank you.

[idle owl]

You're welcome!

[timber mango]

To give you a hint as to what interrupt handling looks like (in Arduino IDE)
@sharp bramble
https://github.com/adafruit/ArduinoCore-samd/blob/master/cores/arduino/WInterrupts.c
maps to:

 $ ls -1 ~/.arduino15/packages/adafruit/hardware/samd/1.0.21/cores/arduino/WInterrupts.c  > /dev/null

Working timer application w/interrupts:
https://github.com/wa1tnr/seesaw_m0/blob/master/m0_timer_ainsuForth/src/app/timer/timer.cpp << *not top-level file; look for the ino file in the base directory for that.

@tidal kiln hey im doing a cleanup of circuitply library if you can hang

(we also fixed all the microphone stuffs)

@tidal kiln did any of your projects use the mic, btw? i can test em

[tidal kiln]

@timber mango don't think so. (at least not any that saw the light of day). let me check though.

[timber mango]

cool

well if you do we have normalized the microphone 😃

[tidal kiln]

not seeing any projects. just the hello_soundsensor

[timber mango]

k ill fix that up too

[tidal kiln]

cool. want me to test anything?

[timber mango]

nah, using travis

[tidal kiln]

k. just let me know if there's anything to check out.

[tidal kiln]

@slender iron what's protocol for adding reviewers (on gh)? i just added circuitpythonlibrarians, but wondering if I'm spamming?

[idle owl]

You're not spamming. That's how it works.

[tidal kiln]

cool. thanks. so it's just a general call out?

[idle owl]

Yep. I've gotten all the emails anyway. If I go to the repo, it will say "Carter has requested your review on this", but when I look at the list, it'll say CP Librarians, and I know I'm a part of that group, and that's how you requested it. It's all part of the process. I'm not getting extra emails because you add me. I think we're all already subscribed to all of these repos anyway.

[umbral dagger]

Trying to add outputting score on the dotstar wing to my game. Fighting CP memory use all the way.

[idle owl]

Sounds about right.

[idle owl]

@tidal kiln 😄 I'm loving your commits.

[tidal kiln]

pancakes!

[idle owl]

Gah I'm getting beat to the merge 😄

[tidal kiln]

gotta be quick, grasshopper

[idle owl]

Apparently. I was digging the board out of a bin.

[timber mango]

zoooom

[idle owl]

yah!

[timber mango]

@tidal kiln i think im nearing the end of the cplay arduino library fix train. some examples in learn need updating but i think they're all very similar.

[timber mango]

@tidal kiln we'll need to update https://caternuson.github.io/Adafruit_CircuitPlayground/

do you want to just move it to the learn guide?

[tidal kiln]

hmmm. yeah. learn would be a better place for all that.

@timber mango so soundSensor has changed?

[timber mango]

yeah

i mean, it works, but i made it signed so 0 is the DC offset. also, it doesnt do what people want which is give SPL. so now there's CircuitPlayground.mic.SoundPressureLevel()

which is what they want

[tidal kiln]

anything else? or is pretty obvious from just looking at source code? that's how i made it before

[timber mango]

thats the only big change

i had to tweak a bunch to get it there 😃

i also doxygen'd the whole thing

so there's detailed function references if desired

[tidal kiln]

so maybe that can be the new API doc ref. and then just get rid of all that static html i created.

[timber mango]

well, a quickstart is still useful

the doxygen is kinda clunky

lets leave it for now

and ill look at cleaning up the library, there's a lot of unnecessary public functions

[tidal kiln]

still want to move that github hosted stuff to learn?

[timber mango]

which onez

oh your text?

yeah if its ez

[tidal kiln]

the api doc, yah

[timber mango]

plz 😃

[tidal kiln]

not too hard, only issue being how best to structure it in learn

[timber mango]

i'd make one long page

with subsections

maybe? or you can just look at the doxygend output

and let me know what you want to change

http://adafruit.github.io/Adafruit_CircuitPlayground/html/class_adafruit___circuit_playground.html

you can link to functions with http://adafruit.github.io/Adafruit_CircuitPlayground/html/class_adafruit___circuit_playground.html#a49e41b3e566d1152daae1caf72d01a22

hm those tags are ugly - why is that :/

[tidal kiln]

yah, i think something in learn would be more user friendly / useful than that

one long page / subsections was what i was thinking too, not sure what else could be done

[timber mango]

i agree

[tidal kiln]

there's a separate learn guide for CPC and CPX, think it could be mirrored between the two?

[timber mango]

yes absolutely!

i will be updating the Xpress guide

the mic was the last, and most annoying, sensor to 'unify'

[tidal kiln]

yah, i'd think so. much different.

when you're done and done, i could add new API doc pages to the CPX guide, then mirror that into the CPC guide

[timber mango]

yay

[tidal kiln]

ha! i'd forgotten all about the PR for hello_captouch

[timber mango]

yah i just did a release, but going forward it will be easier

[slender iron]

<@&356864093652516868> Meeting is normal time (11am Pacific / 2pm Eastern) tomorrow (Monday) here on Discord. @idle owl will be running it since I have a cold. 🤧 I'll be attending and recording though.

[solar whale]

that was the closest thing to chicken soup I could find - feel better.

[idle owl]

@solar whale I think ramen was the closest thing to chicken soup we could get him to eat the other day, so that was appropriate

[manic glacier]

chcbaram

[adafruit/circuitpython] New star added

[crystal pumice]

hey all, i would like to start tinkering with circuitPython, as i understan, the feather M0 or the Metro M0 got the most options in to it(?) y does the metro is more for PRO users? (i am familiar with python and arduino projects)

[stuck elbow]

@crystal pumice the metro is more for people who really like the Arduino form factor, otherwise it's the same as the feather

@crystal pumice just make sure you get the "express" version of the m0 boards — they have extra memory for the filesystem, you will need it

[crystal pumice]

10q 😃

[crystal pumice]

if i would like to add wifi to the M0 express, can it be easily done?

[upbeat plover]

Any plans for a Trinket M0 express?

[stuck elbow]

there were some leeks

@crystal pumice not that I know of

@crystal pumice you can instead use the HUZZAH board, it uses the ESP8266 chip with wifi, but it's a little harder to use, as it doesn't come up as an USB disk, and you have to upload all your files to it using a special program called ampy

[crystal pumice]

@stuck elbow 10x for the answer. what about the ESP32? do u know if its compatible with circuit python?

[stuck elbow]

@crystal pumice it's not at the moment — there is an experimental port of MicroPython for it, but no CircuitPython yet

[crystal pumice]

cheers

[manic glacier]

jerryneedell

[adafruit/circuitpython] Issue opened: #543 ESP8266 build fails for 3\.0

I am unable to build the esp8266 image under 3.0 master - the following error occurs
at the end of the build:

GEN build/genhdr/qstrdefs.generated.h
FREEZE modules
common-hal/microcontroller/Processor.c: In function 'common_hal_mcu_processor_get_uid':
common-hal/microcontroller/Processor.c:43:5: error: implicit declaration of function 'system_get_chip_id' [-Werror=implicit-function-declaration]
     uint32_t id = system_get_chip_id();
     ^
cc1: all warnings being treated as e...

[jovial wind]

@tulip sleet I had a look at the CPX that was being wiped by power interruptions, and it's running CircuitPython 2.2.0

[manic glacier]

jerryneedell

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

It would appear that esp8266/hal-common/Processor.c is missing somthing:

jerryneedell@Ubuntu-Macmini:~/projects/adafruit_github/circuitpython_master/ports/esp8266/common-hal/microcontroller$ diff Processor.c ~/projects/adafruit_github/circuitpython_2x/esp8266/common-hal/microcontroller/
41,47d40
< 
< void common_hal_mcu_processor_get_uid(uint8_t raw_id[]) {
<     uint32_t id = system_get_chip_id();
<     for (int i=0; i<4; i++){
<         raw_id[i] = id >> (i * 8);
<     }
< }

jerryneedell

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

Sorry _ I now see that the update to add the UID was only applied to master, not to 2_x so it just appears that there is some error in the esp8266 configuration.

[manic glacier]

dhalbert

[adafruit/circuitpython] New comment on issue #280: atmel\-samd: Brownout / low battery causes SPI flash erase on Circuit Playground Express

jafine on discord reports a CIRCUITPY erase with 2.2.0, perhaps due to a low battery.

dhalbert

[adafruit/circuitpython] Issue reopened: #280 atmel\-samd: Brownout / low battery causes SPI flash erase on Circuit Playground Express

[tulip sleet]

@jovial wind Could you add a comment to https://github.com/adafruit/circuitpython/issues/280 that describes the scenario in detail? Thanks! If I remember right, it sounds like the battery might have been low to start with. Perhaps the M0 was running OK at the low voltage but the SPI flash chip was not.

[manic glacier]

sommersoft

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

I can try and duplicate later. But, yeah, the UID is only in master..def not in 2.x or any of the 3.0 tags.

@tannewt & @DanHalbert, I can pull the master over to 2.x & 3.0 on my fork and submit a PR...just let me know.

jerryneedell

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

It looks like Travis did not build an esp8266 image when this was merged . Is that intentional?

[solar whale]

@raven canopy the esp8266 issue may be due to how the compiler warnings are set. There several other warning about implicit function definition, but this one causes it to fail.

[stuck elbow]

you don't want this particular warning to be ignored

it would fail anyways at the linker phase

[solar whale]

@stuck elbow there are others in teh esp8288 build and it does compile and link. Why is this one different?

I agree - there should be no warnings, but that is another matter...

hmmm - grep does not ever find this function?? ```jerryneedell@Ubuntu-Macmini:~/projects/adafruit_github/circuitpython_master$ grep -r "system_get_chip_id" *
ports/esp8266/modmachine.c: uint32_t id = system_get_chip_id();
ports/esp8266/common-hal/microcontroller/Processor.c: uint32_t id = system_get_chip_id();
Binary file ports/esp8266/build/modmachine.o matches
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);
ports/esp8266/build/genhdr/qstr.i.last: uint32_t id = system_get_chip_id();
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);
ports/esp8266/build/genhdr/qstr.i.last: uint32_t id = system_get_chip_id();
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void);

[stuck elbow]

right, that's the problem

[solar whale]

yup - thanks

[manic glacier]

jerryneedell

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

looking for the function system_get_chip_id - grep does not find it anywhere in the source tree.

jerryneedell@Ubuntu-Macmini:~/projects/adafruit_github/circuitpython_master$ grep -r "system_get_chip_id" *
ports/esp8266/modmachine.c:    uint32_t id = system_get_chip_id();
ports/esp8266/common-hal/microcontroller/Processor.c:    uint32_t id = system_get_chip_id();
Binary file ports/esp8266/build/modmachine.o matches
ports/esp8266/build/genhdr/qstr.i.last:uint32 system_get_chip_id(void...

sommersoft

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

system_get_chip_id is actually a native Espressif SDK instruction. it is also (and originally) used in modmachine.c for the machine.unique_id function. modmachine.c is in ports/esp8266 folder.

[manic glacier]

jerryneedell

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

It looks like Processor.c needs:

[brisk cairn]

Is there a way for circuit python to detect which board it's running on? For example under Arduino IDE if I want to know if I'm running on a Trinket M0, I can do something like #if defined (ADAFRUIT_TRINKET_M0). Is there something similar in circuit python?

[cunning crypt]

@brisk cairn Yes!

One sec while I double check what it's called

os.uname() pulls the info

https://github.com/Andon-A/CircuitPython-Examples/blob/master/examples/Blink/main.py

I use it there to determine which pin names to use.

[solar whale]

@cunning crypt Nice example! Thanks!

[cunning crypt]

I need to add in the NRF to that one, too.

[manic glacier]

sommersoft

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

Yep. That's the one.

[solar whale]

@brisk cairn you can also access it like this ```>>> import os

os.uname()
(sysname='samd21', nodename='samd21', release='2.2.0', version='2.2.0-5-g544b9e4 on 2018-01-20', machine='Adafruit Feather M0 Adalogger with samd21g18')
os.uname().machine
'Adafruit Feather M0 Adalogger with samd21g18'
"Adalogger" in os.uname().machine
True

[manic glacier]

jerryneedell

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

Do you want to review it and put in a PR or would you like me to create one?

sommersoft

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

Well, it seems that it got dropped off when we were merging the PR. I would have to wait until after work to put in a PR. If you (or anyone) want to submit one before then, by all means proceed.

[brisk cairn]

@jerrn that's exactly what I was looking for. I'm working on my circuit python version of IRLib and all of my examples were designed for use with Circuit Playground Express but I'm now testing it on Trinket M0 and other boards. Previously for input pin I just used board.REMOTEIN) which is the default IR receiver on the CPE. But of course that's not available on other boards. I think I will just create a special module to determine my default input and output pins based on the platform and then that can be imported into all of my example files. Saves duplicating all that code and cluttering up every example program.

[manic glacier]

jerryneedell

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

That what I thought. Since it looks likes some issue with the previous merge, I think it best for you to look it over and make the correction. I'm in no hurry. Let me know if you want me to do anything.

[solar whale]

@brisk cairn all the credit goes to @cunning crypt -- good luck with it.

@tulip sleet @slender iron I noticed that Travis does not do a build for the esp8266 - Is it just not ready for "prime-time" yet? I know there are still some compiler warning for the ESP8266 build - would Travis fail due to those?

[slender iron]

@solar whale we'd like to. its just not set up. the esp sdk stuff isn't simple

[solar whale]

@slender iron ok - understood. Foolishly, I just assumed it was being done already. All the more reason for me to test new releases! Nice to know I can do something Travis can't 😉

[timber lion]

If anyone is curious they're streaming the PyCascades conf talks starting now: https://twitter.com/pycascades/status/955492978046349312

PyCascades (@pycascades)

For those of you who couldn't make it to Vancouver for #pycascades, we are streaming the talks at https://t.co/AcF2BgKz6u 🎥💥

Twitter

Guido is up now

[tidal kiln]

thanks @timber lion
direct link to youtube live stream: https://youtu.be/YBBMI_aAHf8

warts?

[idle owl]

pancakes?

[stuck elbow]

yes, please

[timber lion]

I think I'll miss the 11 unfortunately but quick update from me, wrapped up rfm69 last week, added new boards for rfm69 and rfm9x, updated apds9960 guide, updated thermal printer lib and guide, implemented tsl2591 driver and working on guide update for that one nezt

[manic glacier]

brentru

[adafruit/circuitpython] New comment on pull request #540: issue \#536 CPX change SPI pin designations to externally available pins

@dhalbert Ok - switched pin assignments to mirror your second trial on the REPL.

[tidal kiln]

TSL2591 driver? @timber lion think we can close this?
https://github.com/adafruit/circuitpython/issues/523

[slender iron]

hi hi

I can hear you too

[idle owl]

<@&356864093652516868> Weekly meeting happening in a few minutes if you want to join in

[slender iron]

i can start whenever

actually say something else

ok, now its good 😄

[tidal kiln]

mic check?

[slender iron]

let me know when to start

[tidal kiln]

😦

[slender iron]

can't hear you either

[errant grail]

No mic for me today.

[slender iron]

k going

[raven canopy]

GROUP-hug for keeping up with support here and the forums. @solar whale for staying on the edge of the envelope with testing and finding the little things that trickle past Travis.

[slender iron]

Thanks to @idle owl for running the meeting.
Thanks to @dhalbert for the heap organization idea. Looks very promising.
Thanks to @timber lion for adding RFM69x libraries and new board definitions

maybe hes away

[stuck elbow]

sorry, nothing from me

[slender iron]

hehe, people are typing

[cunning crypt]

Thanks to @raven canopy for the unique board IDs - Those'll be extremely helpful! Also Group Hug, because you all are awesome.

[tidal kiln]

group

[errant grail]

Thanks to @timber lion for the RFM69 library! Will save me a lot of time for my Corrosion Monitor conversion to CPy. Group hug, too!

[timber mango]

working on analog stuff with potentiometers this week.

[raven canopy]

go ahead. I dropped the ball... We'll circle back

[slender iron]

Heap long-lived stuff going well https://www.youtube.com/watch?v=5uBF-bDXs8g

using the memory we have more wisely

[stuck elbow]

nothing from me

[cunning crypt]

I'm going to be updating my "Universal Blink" example to add in the nRF52, and also working on things using the Unique Board ID

[raven canopy]

STATUS: UID needs a little touch up after dropping a line during PR/merge.

VEML6070 library is done, and transferred to Adafruit bundle

[manic glacier]

ladyada

[adafruit/circuitpython] Issue opened: #544 isREPLconnected & isUSBconnected

could be handy to be able to query, from the python side, if USB is enumerated and/or the REPL is opened!

put down for 3.0 but can be long-term

[slender iron]

doing lots of awesome stuff @raven canopy 😃

[raven canopy]

Starting to line up rotary encoder for SAMD21.

[tidal kiln]

• Get Well Soon to Scott
• Added examples to TSL2561 light sensor, ADS1x15 ADC, and OneWire bus
• Have driver for DS2413 gpio breakout, ready to PR it in (Scott - can you create a blank repo I can PR to?)
  https://github.com/adafruit/circuitpython/issues/535
• Had a TSL2591 in the mail, was going to do that next, but Tony beat me to it.

[errant grail]

Ironing out string length algorithms for the StringCar; weather is nicer now. Laying out new Corrosion Monitor to CPy plans. The band has only one more song to record (if you were following along…). This week will find the retro power monitor completed (drilling and machining the enclosure).

[slender iron]

got a link? @umbral dagger

[raven canopy]

@tulip sleet RE: forum re-org... Dilly Dilly!! 🍺

[umbral dagger]

There's a page in the "digital sand" demo guide that is almost ready to go.

[slender iron]

👍

sht31d

[solar whale]

sht31-d

temp/humidity sesnsor

[stuck elbow]

I think there are MicroPython drivers for it

[slender iron]

🤣

[solar whale]

https://www.adafruit.com/product/2857

[slender iron]

it should be a simple release because its just new boards

[solar whale]

@stuck elbow thanks - I'll take a look

[slender iron]

yeah, I think we should do it this week

I can do it

[stuck elbow]

https://github.com/kfricke/micropython-sht31

[slender iron]

nah, no bug fixes

that I know of

yup, all good

pcbs come in today I think

getting close

maybe a couple more days

[timber mango]

hii late but herenow 😃

[idle owl]

Hi! Do you have hug reports?

[timber mango]

hug report to sommersoft for writing up the VEML6070 sensor library!

and dastels for the fun LED libraries 😃

hiiii!

yes!

sorry slow

i commited the ITsyBitsy M0 variant code to 2.2

we'll have that hardware in hand in a week or so

its abotu 20 GPIO pins, + 2MB flash

so bigger than trinket but much smaller than feather

[slender iron]

is it called "express"?

[timber mango]

good for those inbetween projects 😃

yes it is called "express" i renamed that too

wow.

thats my update!

[slender iron]

perfect!

[stuck elbow]

I already have a shield for it on order from oshpark 😃

[formal plover]

I know everyone is talking. Just wanted to stop in and say I miss you all!

[opaque patrol]

should call it itsy blinka

[tidal kiln]

yes

Are we still using / updating the spreadsheet linked from here?
https://github.com/adafruit/circuitpython/issues/269
Or are individual issues good enough for tracking?

[slender iron]

ah, nah I'm behin don it

[timber mango]

^ leeeaky

[slender iron]

I prefer issue now

will update/close that issue

[timber mango]

ISR handling on the horizon/roadmap?

[umbral dagger]

@timber mango Nice!

[slender iron]

its on the radar but not near term

[tidal kiln]

cool. thanks. all i got.

[tulip sleet]

leek soup 🥘

[slender iron]

ISR is complicated

[timber mango]

(sneaky shot 😃

LEEEEEKZ

[formal plover]

leeks

[manic glacier]

jerryneedell

[adafruit/circuitpython] Issue opened: #545 add Driver fo SHT31D Temperature/Humidity Sensor

https://www.adafruit.com/product/2857

[stuck elbow]

why we don't have a leek emoji?

[slender iron]

great job @idle owl ! thanks again for hosting

[tulip sleet]

we have to speak to Unicode about that

[timber mango]

i think tony's out but we have the RFM69 library done and also thermal printer library done and documented

i know it was a common requested set of libraries

[tulip sleet]

Tony is at Pycon Cascades in the northwest

[slender iron]

tony is at pycascades if anyone is there!

[timber mango]

thats it for me

thanks kattni! 😃

[slender iron]

today and tomorrow

[raven canopy]

Rotary encoder: core, or external library? SAMD51 has built in support so magine it will be core...

[timber mango]

i think we'll probably add core support later, using timers rather than the hardware support

[slender iron]

@raven canopy core for SAMD51

interesting

[timber mango]

because the hardware support is fixed on pins. honestly...might as well just use timers for it

[slender iron]

😃

[raven canopy]

Kk

[errant grail]

Thanks!

[slender iron]

thank you!

[timber mango]

thanks everyone!

19:29 UTC

[raven canopy]

Thanks everyone

[tulip sleet]

thanks!!

[slender iron]

recording stopped

thank you thank you

[errant grail]

Get well, @slender iron !

[slender iron]

thanks @errant grail

[umbral dagger]

Awesome job

[timber mango]

Nice job reading all those part numbers! @idle owl

@raven canopy i have a rotary encoder example for samd if you want it

[raven canopy]

@slender iron I'll get with you later about rotary...just looking at it. @idle owl good job!

[idle owl]

Thanks all!

[timber mango]

@solar whale praise in public; criticize in private.

[raven canopy]

@ladyada, I'll take it! I've scoured the Arduino lib and a micropython lib so far. The more the merrier!

[timber mango]

here ya go

And never take the bait.

trinket rotary encoder

[raven canopy]

Thx!

[timber mango]

the 'right' way to do it is all in core, using a timer, any timer really, to check states

but this works surprisingly well 😃

@slender iron It was very late. I was live on discord at the time.

[errant grail]

@slender iron late night here, but not elsewhere in the world.

[timber mango]

I will. Didn't think to do so. @slender iron

[tidal kiln]

encoder counting can be intense, be nice if there could be a background version

[errant grail]

I've stepped in occasionally and try to defuse.

With some folks, it just takes time.

[stuck elbow]

non-native speakers? impossiburu!

[timber mango]

You guys are doing a new thing, moreso that irc.freenode.net which took the first step towards addressing this in a good way.

Be more like Bill and Ted. ;)

[slender iron]

😃

[raven canopy]

That's how I did it once. Use the #code-of-conduct as though it we're Mjölnir

[tidal kiln]

(thor hammer)

[raven canopy]

Thor's Hammer...

[timber mango]

There was a long history of being sardonic in chat and it became a competition (usually friendly). Letting that go is a cultural adjustment. The result is pretty interesting and unexpected by me.

[raven canopy]

Exactly nis. Once you accept a new behavior, that is the standard...

[timber mango]

Also you guys show no enjoyment of having to do your job advising people who cross the line.

[errant grail]

Good discussion. Thanks for the meeting. @idle owl , you did an awesome job.

[tidal kiln]

^ that

[stuck elbow]

thanks all

[idle owl]

Yeah thank you!

[timber mango]

Love those heap movies.

[raven canopy]

Ok. Back to "work" for me. later peoples!

[stuck elbow]

By the way, I know that most people here are on the wrong continent, but there is a going to be a Hackaday confernece soon https://hackaday.com/2018/01/22/hackaday-belgrade-call-for-proposals-now-open/

and the call for papers is open

[tulip sleet]

@stuck elbow are you going to submit uGame or something like that?

[tidal kiln]

months!

[stuck elbow]

@tulip sleet I want to do a talk about the challenges of doing interactive graphics on those cheap displays

[solar whale]

Off for a bit - talk to you later!

[tulip sleet]

https://www.youtube.com/watch?v=l7JHiWISRVU

[timber mango]

https://www.youtube.com/watch?v=UZvmTlAWpaI soundtrack no voices or foley.

Then you can employ SMPTE time signatures to coordinate the heap movie to the music (adding pauses and varying the speed of playback of the visuals, basically). @slender iron
https://en.wikipedia.org/wiki/SMPTE_timecode

[manic glacier]

jerryneedell

[adafruit/circuitpython] New comment on issue #521: on metro\_m4\_express \- CP3\.0 FS error for specific import syntax

Using J-link - so far I have determined that it is ending up in the Hard_Fault_Handler:
line 279 of port.c

 * \brief Default interrupt handler for unused IRQs.
 */
__attribute__((used)) void HardFault_Handler(void)
{
  while (true) {
        asm("");
  }
    for (uint32_t i = 0; i < 100000; i++) {
        asm("noop;");
    }
}

[manic glacier]

jerryneedell

[adafruit/circuitpython] New comment on issue #521: on metro\_m4\_express \- CP3\.0 FS error for specific import syntax

here is a gdb backtrace:

Loading section .text, size 0x2b2a4 lma 0x4000
Loading section .data, size 0x178 lma 0x2f2a4
Start address 0x4000, load size 177180
Transfer rate: 19225 KB/sec, 14765 bytes/write.
(gdb) mon reset
Resetting target
(gdb) c
Continuing.
^C
Program received signal SIGTRAP, Trace/breakpoint trap.
HardFault_Handler () at supervisor/port.c:279
279	        asm("");
(gdb) backtrace
#0  HardFault_Handler () at supervisor/port.c:279
#1  <signal handl...

[idle owl]

Silly question: Mu editor just works, right? As in, do you need to install Python or anything in the background? I ask because I've already had an environment setup on any machine I've installed it on, so I wasn't sure if it was required.

Mac specifically, if that matters.

[tulip sleet]

@idle owl The mu executable on Windows or MacOS contains its own copy of Python.

[idle owl]

@tulip sleet Ok keen, thank you. I have a friend who has decided to get into CP and I want to start him with Mu but realised I wasn't sure if it was self-contained or not

[tulip sleet]

Limor had trouble getting the ordinary .app to work properly so it has that weird terminal-based startup. I think ntoll will look at that and try to fix it.

[idle owl]

Yeah I remember that, only one mac version worked for me, but one was enough 😃

[slender iron]

<@&356864093652516868> Here is the recording from the meeting today. Thanks again to @idle owl for leading it. https://youtu.be/T7yrkrdIxsI

[solar whale]

@stuck elbow Thanks for the sht31 driver reference - I have tested it on an esp8266 with micropython - no problem. Next made a few simple changes and it runs on the esp8266 under CP3.0 but still using the "machine" I2C driver. Next up will be to convert to the bitbangio I2c then to busio on an M0 board. Looks promising though. Nice to be able to follow a progression from micropython to circuitpython.

[umbral dagger]

My CP optimization attempts have prevailed. Added a score output to my dotstar/featherwing game, and after much fiddling got it to fit into memory. I am very interested to see how heap compaction and/or generational GC would help this.

[manic glacier]

jerryneedell

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

FYI - I made the change to add the include and build an image for some other testing. Just for fun, I thought I'd check the UID and this is what I get:

>>> dir(microcontroller)
['__name__', 'cpu', 'delay_us', 'disable_interrupts', 'enable_interrupts', 'on_next_reset', 'reset', 'nvm', 'RunMode', 'Pin', 'pin', 'Processor']
>>> dir(microcontroller.cpu)
['frequency', 'temperature', 'uid']
>>> microcontroller.cpu.uid
bytearray(b'\xe8\x0cL\x00')
>>> microcontro...

sommersoft

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

yeah, that's a correct result. for a human-readable format, import struct and print it using this: struct.unpack('<i', microcontroller.cpu.uid)

we kept the return type as a bytearray so that options were left open to how people wanted to use it.

jerryneedell

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

hmm - whatever it is, it it the same as machine.unique_id()

Adafruit CircuitPython 3.0.0-alpha.1-65-g4bd1d09-dirty on 2018-01-22; ESP module with ESP8266
>>> 
>>> 
>>> 
>>> import microcontroller
>>> microcontroller.cpu.uid
bytearray(b'\xe8\x0cL\x00')
>>> import machine
>>> dir(machine)
['__name__', 'mem8', 'mem16', 'mem32', 'freq', 'reset', 'reset_cause', 'unique_id', 'idle', 'sleep', 'deepsleep', 'disable_irq', 'enable_irq', 'time_pulse_us', 'RTC', 'Timer', 'WDT', 'Pin', 'Sig...

jerryneedell

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

got it - thanks ```>>> import struct

struct.unpack('<i', microcontroller.cpu.uid)
(4984040,)

sommersoft

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

yeah, it uses the exact same function. we added it to microcontroller to keep it standard with the other ports. on the esp8266, the "unique id" is the last 24bits of the MAC address. not quite as unique as other boards..but, its what we got. 😄

sommersoft

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

i'm working on the fix right now...just have to make sure an rebase to avoid further damage. lol

[timber mango]

@slender iron @tulip sleet hey how is random seeded on 2.x

e.g. should we re-seed it with analog readings or something?

[tulip sleet]

@meager fog on esp8266 it's seeded from the HW random # generator. On SAMD it just uses the uptime tick count

[timber lion]

on the esp it can use hardware random, on m0 you need to seed it with something random ideally

ah yeah dan got there

[tulip sleet]

yeah, maybe we could use low-order bits of a few reads of the ADC (like maybe use the internal temp)

[timber lion]

there's a seed function you can call to use anything

i've used light sensor on cpx

let me get a link, the docs weren't great for it in micropython upstream

[upbeat plover]

a combination of both would be cool

light and temp to get the seed

[timber lion]

yeah so import urandom, assuming it's enabled in the port

then call urandom.seed and pass any kind of int or int-like value

https://github.com/adafruit/circuitpython/blob/master/extmod/modurandom.c#L85

i've just used an AnalogIn.value

[tulip sleet]

@meager fog @timber lion SAMD51 has a TRNG: see page 1309 in datasheet

[timber lion]

oh yeah scott mentioned that, should be really nice!

another trick i've used if you're doing like a game and have no good entropy source

use the millis time for when the first button is pressed

not perfect but good 'random-ish' thing

[tidal kiln]

@tulip sleet @timber lion this is on CPX, so SAMD. this code:

import random
from adafruit_circuitplayground.express import cpx

print("press button a")
while not cpx.button_a:
    pass
for _ in range(10):
    print(random.randrange(1,7))

should it print same sequence after each HW reset?

[tulip sleet]

@tidal kiln Does it? In the C code it looks like it inits the seed from the integer value used for time.monotonic (msec ticks).

[tidal kiln]

it does not

when does the seed get called? at import?

[tulip sleet]

@tidal kiln the first time you ask for something random. Code is in https://github.com/adafruit/circuitpython/blob/2.x/shared-module/random/__init__.c

It calls yasmarang(), which checks to see if it's been called before. If not, it sets the seed.

[tidal kiln]

ah. so difference in time to press button = seed

[tulip sleet]

yes

The seed gets set when you call randrange() above

for the first time

[timber mango]

@tidal kiln maybe make a note of that

in the guide - as contrast to arduino

also, could do the same on arduino side - seed with first button press millis()

[tidal kiln]

@tulip sleet ok. second version...

import random
from adafruit_circuitplayground.express import cpx

values = []
for _ in range(10):
    values.append(random.randrange(1,7))
while not cpx.button_a:
    pass
for v in values:
    print(v)

also getting different values after HW reset

[tulip sleet]

that code does NOT wait for button press, so it will not be as random

[tidal kiln]

yep. that's the point. but still different. why/how?

[tulip sleet]

a little surprised since time to first call to randrange() should be about the same. May have to do with USB enumeration time (e.g. host responds sooner or later). But a standalone board not connected to USB might be more consistent in time to first call to randrange()

[tidal kiln]

that makes sense. all that's needed is a difference of 1 time.monotonic unit.

[slender iron]

@tidal kiln it should always be different unless you give the seed explicitly as @timber lion mentioned

[tidal kiln]

@slender iron how though?
just trying to understand the magic, the arduino random is much more not-random and requires seeding. so just curious what CP is doing under the hood.

[slender iron]

@tidal kiln didn't @tulip sleet cover it already?

SAMD21 is seeded by the time of the first call

[tidal kiln]

sounds like it seeds via time.monotonic on first use

[slender iron]

ESP and SAMD51 use the TRNG

you can call os.urandom if you want the hardware generated version when its available

[tidal kiln]

so for the time based seeding, how does code above end up returning different values after each HW reset?

[timber mango]

because they press the button at a different time?

[tidal kiln]

second version doesn't wait for button press

[timber mango]

there might be very slight variations in startup time based on crystal sync setup

[slender iron]

I think dan is right about usb introducing randomness on how fast the code runs

we make no promises about how quickly the python code is run

and so there is error there

[tidal kiln]

k. yep, must be something. kinda neat effect. thanks for info.

[slender iron]

😃

no problem

you could try it off usb with a display to see if its deterministic then

[tidal kiln]

yep

[upbeat plover]

if cpx just light up a pixel for what number

[tidal kiln]

for those still following along at home....

import time
import random
from adafruit_circuitplayground.express import cpx

values = []
for _ in range(10):
    values.append(random.randrange(1,7))
for v in values:
    cpx.pixels.fill((0,0,0))
    cpx.pixels[9] = ( random.randrange(255), 
                      random.randrange(255),
                      random.randrange(255) )
    for i in range(v):
        cpx.pixels[i] = (0,0,255)
    while not cpx.button_a:
        pass
    time.sleep(0.25)

running off batteries....still random after HW reset

although first value is always the same (5)

[opaque patrol]

You could seed off the light sensor or read an analog pin

[tidal kiln]

@opaque patrol yep, this experiment is intentionally checking behavior without a seed, to test actual randomness of random in CP

[opaque patrol]

Okay, I know I tested it back a month or so ago and it didn't seem random

[tidal kiln]

ooo. can you repeat?

[tulip sleet]

could display int(time.monotonic()*1000) in the NeoPixels without USB or on serial with USB (or display last four digits on a 4-digit display or whole number on a SSD1306 or something without USB). That will tell you more directly how repeatable the timing is.

[tidal kiln]

🕳 ..... 🐇 🏃

[tulip sleet]

i was wondering the ultimate point of this exercise

[raven canopy]

careful rabbit! there's a snake in that hole!

[manic glacier]

tannewt

[adafruit/circuitpython] New comment on issue #521: on metro\_m4\_express \- CP3\.0 FS error for specific import syntax

Thanks for the backtrace! I'd suspect the culprit is frame #3. The
HardFault_Handler is where it always ends up.

~Scott

On Mon, Jan 22, 2018 at 1:09 PM jerryneedell [e-mail]
wrote:

here is a gdb backtrace:

Loading section .text, size 0x2b2a4 lma 0x4000
Loading section .data, size 0x178 lma 0x2f2a4
Start address 0x4000, load size 177180
Transfer rate: 19225 KB/sec, 14765 bytes/write.
(gdb) mon reset
Resetting target
(gdb) c
Continuing.
^C
Program re...

[slender iron]

gah!

one yellow block!

[tulip sleet]

very close

[manic glacier]

sommersoft

[adafruit/circuitpython] Pull request opened: #546 Re\-added header inlcusion for system\_get\_chip\_id

During the PR merge for the UID functions, seems like a required header inclusion was dropped off for the ESP8266. This caused a build error. See Issue #543.

Thanks to @jerryneedell for finding and reporting!

sommersoft

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

PR finally submitted. I have ruined my fork's history...to the ProGit book!

Thanks again @jerryneedell for catching this.

[slender iron]

I win! Total free space: 6480 Longest free space: 3408

[solar whale]

I think you should put that on a t-shirt!

[slender iron]

lol, nobody would know

[raven canopy]

put a Waldo in there, and then ask "Where's Blinka?"....
(answer: EVERYWHERE!)

[manic glacier]

dhalbert

[adafruit/circuitpython] Pull request closed: #540 issue \#536 CPX change SPI pin designations to externally available pins

dhalbert

[circuitpython:master] 1 new commit

ab1f133 issue #536 CPX change SPI pin designations to e... - brentru

[sick creek]

where is it in the pic?

[manic glacier]

dhalbert

[adafruit/circuitpython] New comment on issue #521: on metro\_m4\_express \- CP3\.0 FS error for specific import syntax

Reduced this to: from abd import i where
abd/i.py contains:

class I:
    pass

However, if I make the abd folder name a single letter, it sometimes crashes as before, and sometimes works. If I rename it back to abd, it crashes consistently. So seems like it's junk in RAM somewhere. Ugh. Will debug further.

[slender iron]

here is the video for the above: https://youtu.be/S0uEZqxOWOc

[raven canopy]

its got sounds! i was going to mention CGrover should supply the audio, but he had already left voice chat...

[slender iron]

and he read your mind @raven canopy 😉

its from @errant grail

[raven canopy]

yeah, saw that in the YT comments. cheers all around for the "production crew"!! 🎉

[idle owl]

@slender iron Looks amazing!

[raven canopy]

i'd like to figure out the heap on my github fork. git tells me everything is kosher. fetch from both remotes...nothing. github...stuck on "XX commits ahead". If i was sure that upstream history would survive, i would just delete and re-fork.

compare even has me listed on commits i was nowhere near!

[slender iron]

those microbuilder + sommersoft ones are because you rebased the ones from head onto your branch

my guess is that you got the rebase backwards

[raven canopy]

hehe. only me. well, maybe a couple others too.

[slender iron]

did you make a new branch anyways?

[raven canopy]

yeah

[slender iron]

then who cares about master 😃

its just another branch

[raven canopy]

true. i'm vascilating between -d master; -b master....or trying my hand at a revert.

[slender iron]

you can also reset

[raven canopy]

so many options. which to choose... 😝

[slender iron]

I've done git reset --hard adafruit/master

its in my shell history

[raven canopy]

i'll give revert one shot...then i'll probably go hard reset.

[slender iron]

revert creates a new commit I think

reset just changes where the branch points

[raven canopy]

yeah...just remembered that. (after complaining about it what? an hour ago? hahaha)

[slender iron]

😃

[manic glacier]

jafine

[adafruit/circuitpython] New comment on issue #280: atmel\-samd: Brownout / low battery causes SPI flash erase on Circuit Playground Express

I observed this occurring multiple times while we were doing demoing the TR808 drum (with fruit attached). In order to recalibrate, we were un-plugging and re-plugging in a 3.7V lipo battery.One of our kid helpers was leaving the lipo plug loosely in the socket, so that she could quickly disconnect and reconnect. I suspect that it was the quick disconnection and reconnection of power that may have contributed to this. I can re-create the issue by replicating the the loose socket quick switchi...

[stuck elbow]

@tulip sleet Perhaps if there is an explicit command to format the flash, the automatic formatting could be disabled? I noticed that when the flash is not formatted (or otherwise faulty), the serial console still works, so you could issue the formatting command explicitly. Then there would be no problems with brownouts.

[manic glacier]

PublicMakings

[adafruit/circuitpython] New star added

[manic glacier]

frMaciek

[adafruit/circuitpython] New star added

[slender iron]

@stuck elbow that assumes its the autoformatting thats the problem

[timber lion]

day 2 of pycascades stream: https://twitter.com/pycascades/status/955854921827475456 talks start in 10 mins, schedule here: https://www.pycascades.com/schedule/

PyCascades (@pycascades)

Another day another video stream. Today you can follow us live at https://t.co/R4LPkiBpZe 🎥. Thanks @elthenerd!

Twitter

guido's talk yesterday was probably my favorite--really nice overview of lessons learned from python 3

[slender iron]

awesome! I'm gonna watch. the micropython talk is second I believe

the first is about unconscious bias which is important too

[tidal kiln]

youtube link: https://youtu.be/mqMvn6yri4I

[slender iron]

@tulip sleet any progress on the import crash @solar whale found?

just hit it myself

[tulip sleet]

@slender iron just starting on gdb'ing it in earnest. I reviewed the merge and didn't see anything obvious that should affect it. I have to do some heavy tracing. The slightly non-reproducible nature of it (see my last comment in the bug) is too bad.

[slender iron]

k, I'll take a look

ah, my bug is in my code 😃

[solar whale]

@slender iron if it is holding you up, there are pretty easy workaounds.

[slender iron]

I'm on an m0

which it doesn't impact iirc

I just have a different bug

ugh, circular reference

[solar whale]

I never saw it on an M0

[slender iron]

never good when your backtrace has 396 frames

micropython talk is starting!

[sick creek]

@slender iron seems like intel realsense SDK 2.0 have python support

[slender iron]

@fading solstice go ahead and close the lint issue!

great job!

[manic glacier]

tannewt

[adafruit/circuitpython] Pull request closed: #546 Re\-added header inlcusion for system\_get\_chip\_id

tannewt

[circuitpython:master] 1 new commit

fe851fc re-added header inlcusion for system_get_chip_id - sommersoft

[slender iron]

fyi I'm releasing 2.2.1 now

[idle owl]

Oh excellent

[manic glacier]

tannewt

[adafruit/circuitpython] New tag created: 2\.2\.1

mrmcwethy

[adafruit/circuitpython] New comment on issue #475: Lint all the libraries\!

There was a lot of work involved with incorporating pylint into the library build process and it is complete. Thanks everyone!

mrmcwethy

[adafruit/circuitpython] Issue closed: #475 Lint all the libraries\!

[idle owl]

@fading solstice It's the file name of tapodetectsimple.py that has an issue, I just edited the comment to clarify it's not in the file, but I wasn't sure if it would update for you in time.

[manic glacier]

tannewt

[adafruit/circuitpython] New release published: 2\.2\.1

[slender iron]

@tulip sleet what converter do you use to create the release forum post?

[tulip sleet]

http://www.seabreezecomputers.com/html2bbcode/ then I changed "code]" to "b]"

[slender iron]

kk

thanks!

[tulip sleet]

kinda manual, and then making the blog post was another job

[slender iron]

yeah, no worries

[tulip sleet]

blog post, not forum post

[slender iron]

I just did the blog post

I just pull the html out of the github release page

[tulip sleet]

i think I did that, but maybe I had to touch it up a bit, but it's obvious.

[slender iron]

the preview looked ok so I didn't change anything

[idle owl]

@fading solstice Do you want me to do the release on that lib, or do you want to do it?

@tidal kiln travis is unhappy with DS2413. Do I remember you talking about this in an earlier comment thread or is this different?

[tidal kiln]

this is new. the DS2413 repo just went in today. i think it's because i forgot to do a release for the DS2413 repo.

[slender iron]

which travis is unhappy?

the bundle or the repo?

[idle owl]

into the bundle

[tidal kiln]

https://github.com/adafruit/Adafruit_CircuitPython_Bundle/pull/48

[slender iron]

k, perfect. it is supposed to fail if no release has been done

[idle owl]

ok

nice

[slender iron]

after you do the release you can restart it from travis and it'll pick it up

(if you tagged the same commit you wanted to include before.)

[tidal kiln]

release is done. but couldn't figure out the travis restart part.

hmmm. unless....

ah. had to actually sign in to travis. duh.

@idle owl it's happy now

[idle owl]

excellent

merged.

😄 nice

Oops caps.

[half sedge]

Here is a question about future "ItsyBitsy M0 Express" will it have an "enable" pin (like ItsyBitsy 32u4) that would be compatible for use with TPL5111 ?

[tulip sleet]

@half sedge well, take a look at the screen shot of the board in this channel, yesterday 2:25 (Eastern time -- adjust as necessary)

it's "just another M0" board with more pins available, and a one 5V level-shifted output pin, so if the other M0 boards have something compatible, it should too.

[slender iron]

@half sedge It does. I just looked. Its tied to the regulator

but you didn't hear that from me 😉

[half sedge]

Thanks, I inferred that from the name in 2.2.1 release notes. Also I assume express mean not that tiny storage space from Trinked but serious stuff like other express board?

[slender iron]

yup, it has flash too

[manic glacier]

Dynode

[adafruit/circuitpython] New star added

[timber mango]

@half sedge the itsybitsy has EN, yes

@umbral dagger heya for that LSM303 library, did you want to move it to adafruit and we can make it a Real Library? 😃

[umbral dagger]

@timber mango You bet. I'll touch base with @idle owl and @slender iron to make sure everything is ship-shape.

[timber mango]

yay

ok for the code, i put it into our big learning guide repo

that means we can embed it and update the code easily

(rather than have it live directly in the guide!)

https://github.com/adafruit/Adafruit_Learning_System_Guides/blob/master/Digital_Sand_Dotstar_Circuitpython_Edition/digitalsand.py

you can see at the bottom of https://learn.adafruit.com/admin/guides/1909/editor/11424

its an 'embed' type

but if you preview, it pulls the code directly from github!

[umbral dagger]

OK.. I heard something about this but it didn't really make sense until now.

[timber mango]

yeah its basically a way to separate code from learn - to allow for PR's, issues, updates, etc.

but for security we can only embed code from adafruit's github acct (obvs 😃

[umbral dagger]

Neat

[timber mango]

very! ok for the LSM303, yeah its something we've been meaning to implement, one thing to watch for is theres like 3 different LSM303's and they have slightly different names >.<

ok i found a lsm303

@umbral dagger want me to move it now, and i can re-cookie and travis it?

[umbral dagger]

go for it!

woot woot!

[timber mango]

the community is getting to some chips before we do 😃

[umbral dagger]

I'm not surprised by that. You have a lot of sensors.

[tidal kiln]

having pycon livestream pushed to back can cause minor heart attack when they return from break

[umbral dagger]

@tidal kiln The pycascade stream? If so they already released day 1 (not live) .. alas one huge video.

@tidal kiln I've started watching Guido's Python 3 retrospective talk.

[tidal kiln]

yep. i've had today's live stream going on the side.

[timber mango]

@slender iron heya for lib examples they should not be in subfolders right?

https://github.com/adafruit/Adafruit_CircuitPython_LSM303/tree/master/examples -> ill defolder?

[slender iron]

subfolders are ok, lint just needs to be changed to handle them

the featherwing lib has them as well

[timber mango]

hmm! ok up to you 😃

[slender iron]

yeah, either way is fine with me

most don't use them though

[timber mango]

okie doke

[umbral dagger]

I was echoing the arduino library structure

[timber mango]

yah arduino is weird, it requires a subfolder

it doesnt matter, ill keep as is!

(shrug)

[umbral dagger]

It does tend to make it very clear which files you need and which are examples

[timber mango]

true!

ok just tweaking the include names

almsot odne!

will test, i founda breakout

[umbral dagger]

But I'd usually take that further and use an src folder

[timber mango]

lol java programmer here

"lets just have the code ITSELF be FOLDERS"

[umbral dagger]

and C and ruby and haskell...

Although Smalltalk does away with all that folder nonsense 😃

but then, it has its own overwhelming hierarchical structure.

[timber mango]

fixed one typo but i think done now!

@slender iron ah! you cannot run travis on a forked repo?

https://travis-ci.org/adafruit/Adafruit_CircuitPython_LSM303 says no perms?

wait nvmd i have to add librarians. which is odd because im an owner??

[slender iron]

are you logged in as adabot?

[timber mango]

no adafruit

er ladyada

who 'owns' adafruit

[slender iron]

hrm, weird

[timber mango]

could be travis just hiccuped

[slender iron]

ya, there is a delay to it fetching new github data

[tulip sleet]

@slender iron @solar whale fixed the import bug: seems to be an M4 compiler problem. mp_import_from was using char *str = alloca(some_length) to allocate a string on the stack. I changed it to char str[some_length] and it stopped crashing. The asm code is slightly different.

The original bug was apparently that somehow writing to the alloca'd string was smashing the saved pc on the stack so the function return returned to a random place in memory (that didn't exist, which caused a fault).

alloca is used a few other places. I am tempted to change them. Who knows what trouble it's causing on the m4.

[idle owl]

Well done!

[slender iron]

great job @tulip sleet

[raven canopy]

got my replacement Trellis board in today. plan on putting in the leds and starting back to work on a library...but, i need to do some chores. soldering sound more funner...

[timber mango]

@tulip sleet yay! nice work. is this a gcc 7 thing

[slender iron]

@tulip sleet I'll trade you PRs. Not really fair though because mine is much bigger I'm sure

[tulip sleet]

I can't find a report of this - not good. Not sure if I can come up with a test case to report.

@solar whale said it's broken in gcc 6 as well. I haven't confirmed that but will before I try to report it.

[umbral dagger]

I updated the dotstar featherwing guide with the game I showed last week, a more advanced version that required some well-fought optimizations. And a nice video courtesy of my new overhead camera 😃

[timber mango]

@umbral dagger ok i wrapped up the guide and linted it

check it out here! https://github.com/adafruit/Adafruit_CircuitPython_LSM303

[manic glacier]

tannewt

[adafruit/circuitpython] Pull request opened: #547 Introduce a long lived section of the heap\.

This adapts the allocation process to start from either end of the heap
when searching for free space. The default behavior is identical to the
existing behavior where it starts with the lowest block and looks higher.
Now it can also look from the highest block and lower depending on the
long_lived parameter to gc_alloc. As the heap fills, the two sections may
overlap. When they overlap, a collect may be triggered in order to keep
the long lived section compact. However, free space is a...

tannewt

[adafruit/circuitpython] New comment on pull request #547: Introduce a long lived section of the heap\.

I'm still looking into fixing the tests so sit tight.

[timber mango]

@slender iron shoot imforgetting a step - https://travis-ci.org/adafruit/Adafruit_CircuitPython_LSM303 it errors with 'invalid option "--api_key="'

[slender iron]

I gotta run adabot

[timber mango]

oh ok - i made a token

so u can skip that part

[slender iron]

kk

[umbral dagger]

Code in https://github.com/adafruit/Adafruit_Learning_System_Guides/blob/master/Digital_Sand_Dotstar_Circuitpython_Edition/digitalsand.py needs to be tweaked to use adafruit_lsm303 instead of lsm303

[timber mango]

@umbral dagger yah i was gonna do it after this is mpy'd and all that stuffs

[manic glacier]

dhalbert

[adafruit/circuitpython] New comment on pull request #547: Introduce a long lived section of the heap\.

I would like to go first :)

[timber mango]

@slender iron ok let me know when adabot's done and ill keep goin'

[slender iron]

I thought you didn't need it

running now

[timber mango]

oh sorry er - i did the token only!

[slender iron]

on travis? that should be all you need

[timber mango]

!? so weird. i wonder what im doing wrong

[raven canopy]

is he just not starting?

he was super slow the other night..

[timber mango]

nah i think i got my cookiecutters mixed somehow

works now!

[solar whale]

@tulip sleet nice work! I did try it with Gcc 6 and it failed but it would be great for you to reproduce it.

[timber mango]

@umbral dagger whew done! can you tr https://github.com/adafruit/Adafruit_Learning_System_Guides/blob/master/Digital_Sand_Dotstar_Circuitpython_Edition/digitalsand.py

with

https://github.com/adafruit/Adafruit_CircuitPython_LSM303/releases/tag/1.0.1

[manic glacier]

dhalbert

[adafruit/circuitpython] Pull request opened: #548 alloca seems buggy on M4

Fixes #521. The problem appears to be bad compilation of alloca(). alloca() can be replaced with a run-time variable array: that works.

I replaced a few other uses of alloca() that were similar, just in case they might also be problematic. There are some uses where alloca() is not done at the point of declaration, and those are harder to change. Also there are submodules with lots of uses of alloca() that I didn't change.

Eventually I will try to come up with a small test case...

[umbral dagger]

@timber mango Sweet!

@timber mango Thanks

[manic glacier]

jerryneedell

[adafruit/circuitpython] New comment on pull request #548: alloca seems buggy on M4

Great job! - I just tried this PR on my M4 and it works fine in executing the same code that failed for #521

[timber mango]

yay

@tulip sleet dean's going to be doing a light respin of the metro m4 to add an i2s MCK pin

did you get an m4 already with QSPI?

[tulip sleet]

@meager fog not yet - I could wait for the respin

[timber mango]

if its just QSPI we have some now

any other m4 changes we want to make?

[tulip sleet]

not that I know of, I could make an order this week - have some stuff on a shopping list

[manic glacier]

jerryneedell

[adafruit/circuitpython] New comment on issue #543: ESP8266 build fails for 3\.0

Fixed by #546

jerryneedell

[adafruit/circuitpython] Issue closed: #543 ESP8266 build fails for 3\.0

dhalbert

[adafruit/circuitpython] New comment on pull request #548: alloca seems buggy on M4

Unix builds are failing import tests!

[idle owl]

Fancy

[manic glacier]

tannewt

[adafruit/circuitpython] New comment on pull request #548: alloca seems buggy on M4

Weird! There is a Unix test hanging for my pr.
On Tue, Jan 23, 2018 at 6:19 PM Dan Halbert [e-mail]
wrote:

Unix builds are failing import tests!

—
You are receiving this because your review was requested.

Reply to this email directly, view it on GitHub
https://github.com/adafruit/circuitpython/pull/548#issuecomment-359998831,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AADNqUiSP4Y08BaLcZqanZXs532lG9Teks5tNpNFgaJpZM4Rqmpe
.

[idle owl]

@umbral dagger Nice!

[manic glacier]

dhalbert

[adafruit/circuitpython] New comment on pull request #548: alloca seems buggy on M4

Now travis is happy. I had to remove a different alloca() that I had changed to an array. It went out of scope and was destroyed, but the data was needed later. alloca() storage lives until the function returns, so that was wrong in this case.

Original alloca() changed to an array still stands, to fix the bug.

[stuck elbow]

Good news everyone! I just got a photo from the fab!

[manic glacier]

jerryneedell

[adafruit/circuitpython] New comment on pull request #548: alloca seems buggy on M4

I repeated the tests from #521 and this PR still works fine. Thanks.
Do yo have any idea why this cropped up when it did. Was there some change that triggered it or was it just a latent issue and we got lucky?

[hot moon]

@sommersoft#0222 "he"? The adabot is a he? That wouldn't have been my guess. 😃

[stuck elbow]

I think it's an "it"

[manic glacier]

juansoria

[adafruit/circuitpython] New star added

tannewt

[adafruit/circuitpython] Pull request closed: #548 alloca seems buggy on M4

tannewt

[adafruit/circuitpython] Issue closed: #521 on metro\_m4\_express \- CP3\.0 FS error for specific import syntax

tannewt

[circuitpython:master] 3 new commits

4b1e9d8 alloca seems buggy on M4 - dhalbert
1e94c42 undo removal of alloca in builtinimport.c - dhalbert
a5bfbbc Merge pull request #548 from dhalbert/3.0_alloc... - tannewt

caternuson

[adafruit/circuitpython] New comment on issue #535: Add driver for DS2413 GPIO controller

Done.
https://github.com/adafruit/Adafruit_CircuitPython_DS2413

caternuson

[adafruit/circuitpython] Issue closed: #535 Add driver for DS2413 GPIO controller

[raven canopy]

@hot moon I was referring to Travis, the continuous integration bot. I will concede that pronoun assumption is tricky... I may have missed the context; if it were adabot, I would prob use "adabot" or "it". 😁

[manic glacier]

caternuson

[adafruit/circuitpython] New comment on issue #523: Add Driver for the Adafruit TSL2591 High Dynamic Range Digital Light Sensor

OK to close?
https://github.com/adafruit/Adafruit_CircuitPython_TSL2591

[solar whale]

I think Travis is often referred to in a way the adabot will not allow 😉

[manic glacier]

dhalbert

[adafruit/circuitpython] New comment on pull request #548: alloca seems buggy on M4

I did some further testing on this and found that the problem arose when the double-tap stuff was added to the loader map. There's no bug with this:

/* top end of the stack */
_estack = ORIGIN(RAM) + LENGTH(RAM);

The bug appears when _estack is moved down one word to make room for the double-tap check word:

/* top end of the stack */
_estack = ORIGIN(RAM) + LENGTH(RAM) - 4;
_bootloader_dbl_tap = _estack;

I actually just adjusted _estack by -4 and it caused the...

[tulip sleet]

@slender iron I have a non-work mtg in 10 mins but will check with you re 🔼 when that's over (1-2 hours).

[manic glacier]

ladyada

[adafruit/circuitpython] New comment on issue #523: Add Driver for the Adafruit TSL2591 High Dynamic Range Digital Light Sensor

yah lets close - we have a driver :)

ladyada

[adafruit/circuitpython] Issue closed: #523 Add Driver for the Adafruit TSL2591 High Dynamic Range Digital Light Sensor

[slender iron]

@tulip sleet sounds good!

[manic glacier]

jerryneedell

[adafruit/circuitpython] New comment on pull request #548: alloca seems buggy on M4

So - does this imply that there is really no issue with "alloca"? It is just a memory alignment issue?

ladyada

[adafruit/circuitpython] Issue closed: #537 Need Feather M0 RFM69 build with digital pins D3, D4, D8 exposed\.

ladyada

[adafruit/circuitpython] New comment on issue #537: Need Feather M0 RFM69 build with digital pins D3, D4, D8 exposed\.

done in 2.2.1 release - closing!

ladyada

[adafruit/circuitpython] New comment on issue #439: Add support for Feather/FeatherWings w/ RFM95 LoRa

We've also got a binary release for the Feather M0 RFM69's w/CircuitPy 2.2.1 so you can use them as-is for radio projects. We may attack RFM9X next. There's no chance we will fit LoRaWAN in the M0 Py build but we may be able to send/receive LoRa packets (stay tuned! :)

dhalbert

[adafruit/circuitpython] New comment on pull request #548: alloca seems buggy on M4

@jerryneedell I don't know yet. I have something else to do for a while and then will come back to this. It could be that alloca() assumes certain alignment, which isn't true, and the array decl doesn't make that assumption. The generated code does not differ between the working and non-working versions.

[manic glacier]

tannewt

[adafruit/circuitpython] New comment on pull request #548: alloca seems buggy on M4

Assuming word size of 4 bytes was an issue I had with my unix tests as
well. I think making it -8 always would be fine unless the linker script
can actually tell us how big a word is.

On Wed, Jan 24, 2018 at 9:30 AM Dan Halbert [e-mail]
wrote:

@jerryneedell https://github.com/jerryneedell I don't know yet. I have
something else to do for a while and then will come back to this. It could
be that alloca() assumes certain alignment, which isn't true, and the array
d...

[half sedge]

@stuck elbow What is this gaming device? Sound like a Pokitto (for witch somebody already compiled circuit python).

[slender iron]

FYI, <@&356864093652516868> we're going to disable rebase and squash merges on our repos. Regular merges should be used going forwards. (I've been using them more than anyone else I think.)

[manic glacier]

tannewt

[adafruit/circuitpython] New comment on pull request #547: Introduce a long lived section of the heap\.

Ok, this is ready for review.

[stuck elbow]

@half sedge yes, it's very similar to Pokitto or Gamebuino, except this one runs CircuitPython

@half sedge it's something I'm working on after hours, the project page is at https://hackaday.io/project/27629

[idle owl]

@slender iron sounds like a plan.

[half sedge]

@stuck elbow You may want to check this https://talk.pokitto.com/t/wiki-draft-introduction-to-micropython-on-pokitto-hw/597
I am not an M0 expert, but it felt like Pokitto is also a M0. Is that possible?

[sick creek]

who was the jupyter guru?

jupyter notebooks

here is something who like that tool https://notebooks.azure.com/ as this is powered by it

[stuck elbow]

@half sedge it's also an ARM Cortex M0, but not Atmel's, so a special port is needed. µGame uses SAMD21 just like the Feather M0 boards (and also like Gamebuino)

[manic glacier]

dhalbert

[adafruit/circuitpython] New comment on pull request #547: Introduce a long lived section of the heap\.

line 150 and 152 are both setting MP_STATE_MEM(gc_last_free_atb_index), so line 150 is wrong or redundant?

dhalbert

[adafruit/circuitpython] New comment on pull request #547: Introduce a long lived section of the heap\.

What's the 4? Is that number of bytes? could it be 8 on 64-bit-word impls?

dhalbert

[adafruit/circuitpython] New comment on pull request #547: Introduce a long lived section of the heap\.

Is this still a question?

[idle owl]

@slender iron Does the motor init PR still need to be tested?

[slender iron]

the featherwing one? I tested it before I made the PR

[idle owl]

ok.

@slender iron merged and released.

[slender iron]

thanks!

almost through all of the open lib issues

[idle owl]

I see that 😃

@slender iron You and I had talked about making a separate color/animations lib for lights. Do you want me to comment on that issue in NeoPIxel?

I still have it on my list from ages ago.

[slender iron]

sure, coordinating with @dry vortex on it would be good too

[idle owl]

Should I tag them in my comment as well then?

[slender iron]

sure!

[idle owl]

keen!

[slender iron]

k, lunch time

[bronze geyser]

i've loaded feather_m0_rfm69_firmware.bin along w/ adafruit_bus_device and adafruit_rfm69.mpy. Simpletest.py runs. I'm able to get stuff like Temperature, e.g.: '''Temperature: 27.0C
Frequency: 915.0mhz
Bit rate: 250.0kbit/s
Frequency deviation: 250000.0hz
'''
But not sending packets (i.e.: both sides: Received nothing! Listening again....)...just looking for some guidance on best way to debug. Thank you.

[manic glacier]

dhalbert

[adafruit/circuitpython] Issue opened: #549 atmel\-samd: Align stack on 8\-byte boundary

The stack should be aligned on an 8-byte boundary, according to the ARM procedure calling standard (AAPCS). A 4-byte boundary might work, but the interrupt controller expects an 8-byte-boundary, and compilers will also expect this.

Right now _estack is on a 4-byte, but not an 8-byte boundary, due to the introduction of _bootloader_dbl_tap. This appears to have caused issue #521, which was fixed by PR #548, which avoids using alloca() in mp_import_from(). My guess is that gcc assume...

dhalbert

[adafruit/circuitpython] Issue opened: #550 use correct linker maps for SPI flash boards

In 3.0, boards with external flash are still using linker maps for internal flash. Correct the value of LD_FILE in the mpconfigboard.mk files.

jerryneedell

[adafruit/circuitpython] New comment on issue #549: atmel\-samd: Align stack on 8\-byte boundary

Nice! This makes a lot of sense. I suppose this is better than finding a compiler bug!! At least you can fix this.

[idle owl]

Great job, @solar whale! https://learn.adafruit.com/micropython-hardware-sd-cards/code-storage-on-sd-card

[stuck elbow]

8GB of code!

[solar whale]

give you a use for all those small SD cards that are laying around.

[stuck elbow]

I wish I could buy 1MB microsd somewhere for cheap

they would be perfect for cartridges to my game thingy

[solar whale]

I have several 4Gbyte cards from early Raspberry Pi that were just sitting around.

[idle owl]

I found a 2GB normal SD sitting around from an early camera or something. It was still in the sealed package so I must have thought that was a ridiculous size even at the time and purchased a larger one.

[stuck elbow]

I have a bunch of those 2GB cards

but they are still expensive, relatively speaking

[idle owl]

Looking it up, I'm not sure they were ever available smaller than 2GB.

Nope, I'm wrong. Read it again. 32MB looks like where it started.

[stuck elbow]

yeah, I remember some 300Mb usb disk drives

[idle owl]

Yeah same. And they were ridiculously expensive.

[manic glacier]

dhalbert

[adafruit/circuitpython] New comment on issue #550: use correct linker maps for SPI flash boards

This was actually already fixed. I was looking at an old commit when I thought I saw this problem.

dhalbert

[adafruit/circuitpython] Issue closed: #550 use correct linker maps for SPI flash boards

tannewt

[adafruit/circuitpython] New comment on pull request #547: Introduce a long lived section of the heap\.

150 was wrong. Good catch!

tannewt

[adafruit/circuitpython] New comment on pull request #547: Introduce a long lived section of the heap\.

I'm still unsure about it but the comment isn't useful so I removed it.

tannewt

[adafruit/circuitpython] New comment on pull request #547: Introduce a long lived section of the heap\.

Its the number of pointers stored in mp_obj_fun_bc_t before the extra_args array. Is there another way to get the array length?

dhalbert

[adafruit/circuitpython] New comment on pull request #547: Introduce a long lived section of the heap\.

The struct defn is:

typedef struct _mp_obj_fun_bc_t {
    mp_obj_base_t base;
    mp_obj_dict_t *globals;         // the context within which this function was defined
    const byte *bytecode;           // bytecode for the function
    const mp_uint_t *const_table;   // constant table
    // the following extra_args array is allocated space to take (in order):
    //  - values of positional default args (if any)
    //  - a single slot for default kw args dict (if it has them)
...

dhalbert

[adafruit/circuitpython] New comment on pull request #547: Introduce a long lived section of the heap\.

I guess it's allocated separately and assigned to there?

[tulip sleet]

@idle owl @solar whale I have a 16MB "MMC" card (SD card more or less, I think: same shape) I got with a camera which I used when testing the Windows delayed-write issues. Also have a 10MB CF card from an even earlier camera.

[idle owl]

There were a bunch of revision form factors between the main ones that were very short lived, but were produced

[solar whale]

I had a box with old cards in it, but my recent workspace move has made it dissappear! I think there were some very small micro SD cards in it.

[tidal kiln]

remembers having a camera that used 3.5" floppy disks

[solar whale]

Don’t get me started. I used 8” floppies....

[timber mango]

you have not lived unless you edited paper tape with scissors 😃

[idle owl]

Always mark the stack of punch cards with a marker on the edges!

[tight flax]

oh, I remember punch cards, and teletype with long reels of paper tape.

With the feeder we had, you could set it to hold off between tapes, and it would just transmit the “all holes punched” signal until you fed the new tape in.

So you didn’t have to cut and glue or anything

But if you didn’t catch end-of-tape in time, it would disconnect. I remember racing to the Telex machine more than once.

[stuck elbow]

I used to do ascii-art on a telex as a kid

I still have the tapes somewhere

[tight flax]

We used to sneak pics like that back and forth over the telex while waiting for a document to send

Those were awesome

[manic glacier]

dhalbert

[adafruit/circuitpython] Pull request closed: #547 Introduce a long lived section of the heap\.

dhalbert

[circuitpython:master] 6 new commits

6560596 Switch to m_malloc_maybe and m_free to reduce o... - tannewt
56bd078 Speed up qstr loading by using the stack to sto... - tannewt
416abe3 Introduce a long lived section of the heap. - tannewt
da330f0 Polish up the heap analysis script and make it ... - tannewt
aa0ce98 Fix the initial state and polish a couple comme... - tannewt

dhalbert

[adafruit/circuitpython] Pull request opened: #551 3\.0 double word stack align

Undo #548 with a revert and replace with a better fix. Fixes #549: align stack on a double-word (8 byte boundary).

[manic glacier]

tannewt

[adafruit/circuitpython] Pull request opened: #552 Make parsing more memory flexible\.

The parser attempts to allocate two large (~512 byte) chunks up
front. If it couldn't, then it would error out. This change will
cause it to try allocating half the previous attempt until its down
to two copies. This is ok upfront because later code checks bounds
and tries to extend the allocation if needed.

Fixes #533

tannewt

[adafruit/circuitpython] Issue closed: #549 atmel\-samd: Align stack on 8\-byte boundary

tannewt

[adafruit/circuitpython] Pull request closed: #551 3\.0 double word stack align

tannewt

[circuitpython:master] 3 new commits

302ad40 Revert "alloca seems buggy on M4" - dhalbert
efbf082 make _estack be double-word aligned - dhalbert
a1254f1 Merge pull request #551 from dhalbert/3.0_doubl... - tannewt

[manic glacier]

dhalbert

[adafruit/circuitpython] Issue closed: #533 Improve the parsers resilience to low memory conditions

dhalbert

[adafruit/circuitpython] Pull request closed: #552 Make parsing more memory flexible\.

dhalbert

[circuitpython:master] 2 new commits

dff7445 Make parsing more memory flexible. - tannewt
271c4be Merge pull request #552 from tannewt/flexible_p... - dhalbert

[timber mango]

We used broken pieces of brick to write simple symbols such as X and .. well, X was pretty much it.

On other masonry, such as concrete.

[timber mango]

One day my Dad was looking at his 80286-based computer -- it was a Compaq, and the third iteration of personal computer ownership for him -- and, with it's 40 megabyte hard disk drive, he smiled...

Does the adafruit trinket work with LCDs neopixels and micro servos

A servo needs one PWM pin (that's a YES) (set to 50Hz and 3-10 percent duty cycle); an LCD needs at least two pins, SDA and SCL (if I2C-enabled) so that's a YES -- and NeoPixel is a YES as I have one running an 8x strip of them.

Also I answered as if you'd asked about Trinket M0. There are earlier Trinket models I don't know about.

Okay thanks all I need to know

Ref.
Sales sheet https://www.adafruit.com/product/3500
Pinout https://learn.adafruit.com/assets/49778
Example code https://learn.adafruit.com/adafruit-trinket-m0-circuitpython-arduino/circuitpython-pwm

[last drum]

I'm looking for sample code to use one the pads as an hardware interrupt input. Is there such a thing?

[timber mango]

The hardware supports it but you're looking at the Arduino IDE (C++) to realize it in software.

[last drum]

is it on the 'todo' list for the devs?

[timber mango]

They're using words like 'long term' so I'm not expecting it for first half of 2018. 2019ish sounds likely. Also 'it is not exposed' was mentioned, which leads me to believe a motivated person could find uses already implemented.
Some kind of a fig-leaf doctrine going on, there.

[raven canopy]

@last drum ISRs are a frequent topic. And yes, the team is most likely looking at it as a "long term" project, as @timber mango said. No details on when/if to expect it. MicroPython has limited ISRs, but they're on different chipsets. For a read on the pitfalls that come with ISRs on MicroPython, here is a good read: https://micropython.org/resources/docs/en/latest/wipy/reference/isr_rules.html#

[timber mango]

There aren't a whole lot of on-chip interactive environments, where all you need on the road is a serial terminal and the target board. Then there's the ease of .UF2 drag and drop software-installation, cross-platformed to three distinct environments (Makecode, CircuitPython; Arduino IDE).
So there are compelling reasons to be excited about the current state of the hardware and the software that runs on it, and a rather wide variety of hardware form-factors to choose from, all based in SAMD21 (the ARM M0+ microcontroller).

[solar whale]

@tulip sleet Have you checked on how nano does file saves on linux boxes? Is is considered "safe" for CircuitPython? It is not in the list, but I was curious if there are know issues with it.

[tulip sleet]

@solar whale I'd have to find the source code. I've done that for some other editors. Wouldn't take too long.

[solar whale]

@tulip sleet no rush - I can look as well, but you know better what you are looking for.

[tulip sleet]

@solar whale no it doesn't do any fsync-ing or similar. Looked in https://github.com/dtrebbien/nano, esp https://github.com/dtrebbien/nano/blob/master/src/files.c. I did clone and search the whole tree too.

not a good choice of editor for us

[solar whale]

Good to know - thanks!