right?
#circuitpython-dev
exotic pumice
tulip sleet
what is it supposed to be?!
tulip sleet
you mean they only have one clock for both?
exotic pumice
yeah
so if you create an enum out of that chart without reading all the indexes, you end up off by one and you turn on the wrong clock instead of the sercom clock
tulip sleet
I think you should file an issue with MicroChip about that, to note it in the doc.
exotic pumice
haha, there are several issues I could file with microchip
tulip sleet
I'm just looking to see if it's a typo or something
exotic pumice
they state that divsel makes the divider 2^(n+1) when in fact it's 2*(n+1)
I've found a few issues in their SVD as well
umbral dagger
Any objections to me calling my CP version of Lispy "CircuitLisp"?
tulip sleet
They do respond to filed cases via their issues/case mechanism. At least it will be put on someone's TODO list.
@exotic pumice
Note entry 26. The entries for 9 should be like that.
@umbral dagger I'd say pass it by pt.
exotic pumice
or like entry 3
tulip sleet
yah
exotic pumice
where is this issue tracker?
tulip sleet
hold on...
@exotic pumice https://microchip.secure.force.com/microchipknowledge/articles/en_US/FAQ/How-to-submit-a-case/?l=en_US&fs=Search&pn=1 you'll need to register for an account
exotic pumice
alright, I'll get on that
Hello,
I spent the last two months trying to track down a bug that ultimately ended up being due to the confusing way you have listed TC0 and TC1 in the PCHCTRLm table, table 14-9. You have two entries in the table with the same index (9), but no visual indication that they are linked. This caused me to write an enum starting at 0, going along every row of the table, which led to an off by one error, because I didn't see TC0 and TC1 were actually the same index. I recommend you change the way you have noted TC0 and TC1 to something more like entry 3 or entry 25, so it is more visually clear that they are the same index. I'm hoping an improvement to this documentation can prevent others from making the same mistake I did.
Cheers,
Paul Sajna
submitted one for divsel, now I gotta remember all the SVD issues I found
maybe I'll just send them a diff as a .patch file
meager fog
@exotic pumice wow - nice find
there are some typos in ASF and they are Really Exciting :/
solar whale
@slender iron FYI -- I was able to get the displaytest you posted yesterday running with these two changes -- do they make sesnse?
meager fog
@gilded cradle heya do you need any help with seting up ESPSPI? ive been swamped π
solar whale
@slender iron second change
@meager fog I think she has it running -- she tested some examples I posted in a PR to the repo
exotic pumice
I filed 3 issues.
marble hornet
@slender iron any naming preference on the shape lib?
meager fog
@solar whale cool - yeah i commented that maybe a wifi manager could be useful. i dont think ill get to writing one soon - i have some socket optimizations to work on which are weedy
but if you want, you could gank some of my code from here https://github.com/ladyada/Adafruit_CircuitPython_PyPortal/blob/master/adafruit_pyportal.py
to make a happy wrapper that is optimized for requests which, honestly, is what everyone wants
like you'd call get or post directly on the object and it would ferry that to requests for you
then you just init it, and call get/post to fetch data. then on each call it would also check connection n all that
read settings.py n all that
maybe also expose ping
i like using the neopixel for status too, blinks when xfering data, red when connecting to wifi, etc
solar whale
OK - thanks -- will take a look -- lots of new stuff for me so it may not be fast...
meager fog
any suggestions?
you dont have to write it all, i can probably do some π
just figure it would be nice to not replicate the access point code
marble hornet
and a shapes ever not in a tilegrid?
solar whale
@meager fog OK -- took a few minutes to sink in , but I think I see waht you are looking for. sounds likea good way to go. I'll see what I can do - should have some time this weekend
or if @gilded cradle would prefer to take it on, thats fine with me
slender iron
@solar whale ya, those fixes look good
@marble hornet how about "Adafruit_CircuitPython_Display_Shape"?
and ya, I think all in a tilegrid to start so that the api can include color
marble hornet
https://github.com/TG-Techie/TG_Display_Shape mind think alike
give me a min to commit my local code
i haz more questions, do you have a few min?
marble hornet
pushed!
okay two things: does adafruit mind forking from my repo? (if not this is me giving express permission to gank the code as long as it is attributed)
idle owl
@marble hornet We have forked other repos previously, but it's usually easier to create a new repo. You have to put all the code into cookiecutter and deal with all of that anyway. And of course you would get attribution π
gilded cradle
Hi @meager fog, yeah I got it working. @solar whale, yeah I could probably take that on. Seems pretty straight forward and is definitely in my wheelhouse.
solar whale
@gilded cradle Great! Go for it -- let me know if you need any help or testing.
gilded cradle
Ok. I'll have to wait until later tonight to work on it, but I'll let you know if I have any questions.
solar whale
I would not have done anthing until Friday at the soonest.
gilded cradle
Yeah, I'm trying to juggle a bunch of stuff with moving and all, so I'll need to sneak in a little development here and there when I can.
solar whale
Life happens π Good luck moving?
lone sandal
gilded cradle
Thanks @solar whale
solar whale
@tulip sleet ran into an interesting issue -- I am testingthe adafruit_featherwing library alphanum or sevensegment functions and they work fine on a feather_m4_express, but they put the feather_nrf52840 (and argon) into safe mode -- I am running current master on all boards
gilded cradle
They use the Adafruit_CircuitPython_HT16K33 library. Could you try the example in that?
solar whale
actually they crash the nrf boards -- drop USB and end up in safe mode
tulip sleet
could you figure out the simplest example and file an issue? They are I2C, right?
solar whale
yes -- working on a simple example -- will file an issue -- just wanted to give a heads up. Yes I2C
tulip sleet
we haven't changed anything about I2C recently, so hmm
solar whale
I've heard that before π
tulip sleet
i'll set it up with a doubler or something to debug. It might be due to I2C hogging the CPU so the BLE softdevice gets mad, maybe, maybe
solar whale
OK -- I'm just running the "simpletest" for the alphanum or sevensegment (in latest pr). I'll try the ht16k33 test now
ht16k33 test works OK on the feather_nrf52840
gilded cradle
Oh, that's good news. Can you tell which functionality on the FeatherWing library simpletest it fails on? I usually try and have it do a thorough range of stuff.
solar whale
not yet -- it fails very quickly at the start and the terminal session closes so I loose any messages. working on a test
gilded cradle
Ok. I was thinking of commenting out most of the stuff and slowly uncommenting until it fails.
solar whale
Ill first jsut run in manually in the REPL
gilded cradle
That works too
solar whale
very strange -- the alphanum works fine when I run the commands normally fial the REPL
gilded cradle
Maybe it's running stuff too rapidly that causes it?
gilded cradle
kk
solar whale
it crashes on the import of the featherwing lib.
this kills it ```print("import sleep")
from time import sleep
sleep(1)
print("import featherwing")
from adafruit_featherwing import alphanum_featherwing
Ok, that's odd. I'll take a closer look at the code. I just ordered an NRF52840 so I can use that for testing in the future.
solar whale
from adafruit_featherwing import alphanum_featherwing works via REPL but crashes if in script. just the one line
gilded cradle
I wonder if we can change it to import alphanum_featherwing and then in the instantiation change it to alphanum_featherwing.AlphaNumFeatherWing(
solar whale
@tulip sleet see above -- one line kills it -- shpuld I file the issue in circuitpython or featherwing?
gilded cradle
@solar whale, could you try different versions of CP?
solar whale
yes, will try that
solar whale
not sure if 3 support the nrf boards
gilded cradle
Oh, good point
solar whale
I try earlier version of 4.
gilded cradle
Ok. If we can get it to not fail then it may be a CP bug.
tulip sleet
@solar whale @gilded cradle if you can, try beta.1 or alpha.5 on the nRF too, since there have been significant changes since then
solar whale
it fails on beta2 and on beta0
tulip sleet
beta.2 changed how the BLE and tinyusb stuff were set up. It's a significant change for the nRF52840
gilded cradle
Yeah, I think I tried it on all betas while developing with the Feather M4 Express
However, I'll try it on the nrf board when that comes in
solar whale
I have to rebuild before beta0 since there are not .uf2s --- but since it work in REPL -- hard to thin whats going on.
but since it does work in the RPL, hard to blame the library... I'll open an issue in CP.
gilded cradle
Well, what probably happens is that when it's imported, it cascades other imports within those libraries, so maybe we need to dig a little deeper.
However, I'm not sure why it wouldn't fail in the REPL in that case, but worth looking at.
raven canopy
i'm trying to find it, but i remember an issue with periph calls too early after reset... can't remember if it was in the GH issues or on the forum.
manic glacierBOT
solar whale
I have to go for a few hours -- will dig more later.
gilded cradle
Ok, thanks for your work with looking into that Jerry
solar whale
it would be nice if someone can reproduce the issue
raven canopy
"You build it, we'll break it!" - Jerryn Industries. π Thanks @solar whale. if i had any of my particle boards CP'd, I would try. prob won't have any cycles tonight...
uh-oh...
/usr/bin/../lib/gcc/arm-none-eabi/7.3.1/../../../../arm-none-eabi/bin/ld: build-itsybitsy_m0_express/firmware.elf section `.text' will not fit in region `FLASH'
/usr/bin/../lib/gcc/arm-none-eabi/7.3.1/../../../../arm-none-eabi/bin/ld: region `FLASH' overflowed by 7380 bytes
collect2: error: ld returned 1 exit status
Makefile:446: recipe for target 'build-itsybitsy_m0_express/firmware.elf' failed
make: *** [build-itsybitsy_m0_express/firmware.elf] Error 1
may have to try and update the branch to current, or abandon M0. π¬
raven canopy
oh...that's even more perplexing. π€
gilded cradle
What do you mean by manually imported script?
gilded cradle
Huh, I'll have to try that.
solar whale
REPL
gilded cradle
Ah, ok. That makes sense then.
gilded cradle
Yeah, I just was asking how jerry did it.
raven canopy
its a nice little technique. reduces having to keep track of all those code/main .py's. π
solar whale
I just copy the file to the.board then import.
gilded cradle
I hadn't thought about doing that, but I probably will more often now.
raven canopy
whats a finished project? never experienced that... π€£
gilded cradle
I use it for dev work because I can just save it and it reloads
solar whale
That scares me π
gilded cradle
Well, it sure speeds things up. However, I can see it being a problem if it puts your board in a looping state.
In that case, it may be good to place a 2 second delay before code starts executing so you have a chance to put it in REPL mode.
solar whale
but now you can force boot to safe mode so you can remove the file
gilded cradle
How do you do that?
solar whale
slow double tap on RESET button about .5 sec -- instead of going to bootloader - it goes to safe mode -- yellow neopixel
gilded cradle
Oh, thanks. I didn't know about that.
solar whale
i have never used until just now it works!
gilded cradle
π
solar whale
more like 1 second between tap
gilded cradle
Ah, ok
solar whale
I thought I was leaving ... soon
gilded cradle
π
solar whale
looking at the library code -- it ia hard to see a cause, but as I have learned, when code does not work, stop focusing on waht it is supposed to do and focus on what it is doing....
gilded cradle
Yeah
raven canopy
hmm. time to ask myself this question again: to git reset --hard or wade through a painful merge/rebase. reset and re-glue wins...
solar whale
I'm fond of git pull --rebase upstream master git push --force-with-lease origin master
raven canopy
i guess i could give that a shot. i am way off from current, and have changes peppered throughout. thanks @solar whale
i've backed up the "difficult" stuff, so if i lose it all the easy things won't take to long.
solar whale
at least it's a known starting point. just save your changes!!
yay -- finally time to go -- I'm hungry -- back later.
raven canopy
windows and symlinks. π‘
wild pasture
@tulip sleet Howdy, you were awesome helping me with some tips on MemoryError yesterday. I uploaded my code here: https://github.com/steclarkdev/BeeLamp/blob/MemoryErrorCode/code.py
@tulip sleet Really, I was hoping there was a comprehensive guide on memory profiling, optimization etc. I did my due diligence but only found the one page on expectations that mentioned around 250 lines of code. CPX board
@tulip sleet Idk if there's coding tips somewhere or a guide to "Just skip python and move to c code if you...".
tulip sleet
@wild pasture i'm in a mtg but will get back to you!
wild pasture
@tulip sleet No rush! Sorry, i had to bail yest bc I have a 13 week pregnant gf who needed dinner ASAP lol. Ty, appreciate it!
slender iron
@wild pasture I did a memory deep dive a couple weeks ago if you want to jump in the deep end: https://www.youtube.com/watch?v=baa5ILZTRkQ
umbral dagger
No ure module in the Feather nRF52840 build?
slender iron
Β―_(γ)_/Β―
umbral dagger
It's there on the grand central and feather m4 & itsy bitsy m4
tulip sleet
@umbral dagger, that's an accident; could you file an issue? Is json there also?
is it re instead of ure?
wild pasture
@slender iron will do. Any self-service i can do i will. Appreciate it!
tulip sleet
@slender iron ohmygoodness, the nrf port is actually using MICROPY_OBJ_REPR_A, and atmel-samd is using REPR_C. REPR_A uses heap-allocated floats (!) (which means two more bits of accuracy, but more ram)
like I said, I'm going to refactor the includes so these are all consistent
slender iron
interesting! I'm considering boxing in 2-tuple of small int
really handy for position tuples
tulip sleet
you could overlay that with the 31-bit ints, I guess - no need to actually make it identifiable as a 2-tuple
atmel:
>>> 1/3
0.333333
nrf:
>>> 1/3
0.3333333
we're not going to keep this
raven canopy
@slender iron @tulip sleet i'm trying to get FrequencyIn back to an active review, but in seeing that we're out of memory on M0 express, I'm guessing that I should just scrap M0 support?
slender iron
@raven canopy ya, maybe make a frequencyio for it
you can leave any m0 code in that you have
raven canopy
π€ yeah, that would probably be the easiest way. seeing as how we're trying to get out of the #ifdef quagmire.
slender iron
yup yup
exotic pumice
haha pound-define heck
raven canopy
good thing i'm re-gluing anyway... haha
exotic pumice
I'd be happy to help review if you like
seeing as I know so much about clocks now π
umbral dagger
I just submitted a very (IMO) interesting guide and code. π
exotic pumice
lispy?
umbral dagger
Now officially CircuitScheme
exotic pumice
cool
umbral dagger
It has a ways to go, but it's a good start. A better REPL is next on my list for it. I expect it to be an ongoing project for a while.
exotic pumice
saaaaaaaaaame. I just "finished" atsamd-rs for samd51 but it has a ways to go for things like the dac, adc, dma, rng, aes, etc.
If I want to build up abstractions on all those
solar whale
@gilded cradle I wonβt be able to do any more with the FeatherWing lib on the nRF52840 tonight. May be Friday before I can spend much time on it. In the meantime can someone with an nRF52840 try to reproduce the issue?
https://github.com/adafruit/circuitpython/issues/1545
gilded cradle
Ok @solar whale . If somebody else isn't able to, I should get my board and be able to take over on Friday.
solar whale
Great! Iβll be able to dig into it as well. Iβm really curious what is going on.
tulip sleet
i'll look at this also, but prob not tonight
exotic pumice
oh btw @raven canopy , I ended up needing a C version of clock_out for my debugging, might help you out in creating your own test frequencies
#include "sam.h"
int main(void)
{
GCLK->GENCTRL[2].reg =
GCLK_GENCTRL_SRC(GCLK_GENCTRL_SRC_DFLL_Val) |
GCLK_GENCTRL_GENEN |
GCLK_GENCTRL_DIV(4u) |
GCLK_GENCTRL_OE;
PORT->Group[0].DIRSET.reg=PORT_PA16;
PORT->Group[0].PINCFG[PIN_PA16].bit.PMUXEN = 1;
PORT->Group[0].PMUX[PIN_PA16 >> 1].reg |= GPIO_PIN_FUNCTION_M;
while (1) {}
}
indentation is hard
raven canopy
hehe. yeah, especially in the code blocks. the kerning there is...bleh.
exotic pumice
that's gclk2, 12MHz, PA16 (metro m4 d13)
you have to pick a gclk matching your pin
it's under pin function m on the datasheet
or h on '21
in which case you also have to change the 12 up there
I wrote this in a hurry π
there
main meteor
@slender iron looks from that deep dive video like you're a fellow Graphviz aficionado.
I regularly abuse Graphviz
slender iron
π it worked well enough
exotic pumice
@tulip sleet We did it!
exotic pumice
@raven canopy have you looked into the freqm peripheral for frequency in on D51?
I guess you probably want something that works on 21 and nRF and such
tulip sleet
@exotic pumice Yay!
manic glacierBOT
raven canopy
@exotic pumice I have. But FREQM is only for measuring internal clocks. I actually used it in a previous iteration, to help adjust readings due to open-loop DFLL drifting.
raven canopy
@exotic pumice here is the since reverted implementation of FREQM I did: https://github.com/adafruit/samd-peripherals/commit/51784042a567d703e0a1c2f138875d4d203f2bb2
exotic pumice
@raven canopy It looked like you could feed in external clocks on gclk_io, the way I read it
umbral dagger
@tulip sleet Let me know when ure is available in the '840 build and I'll build/test it against CircuitScheme.
raven canopy
@exotic pumice well, yeah, that might actually work. Could be an interesting dance with configuring the REF_CLK though (lower-but-not-too-low frequency, adequate REFNUM). Not that using the TC hasn't been a dance. Haha
And thanks for pointing that out. My inexperience let me look past that as an option.
lime trellis
@exotic pumice just read your comment about atsamd-rs. VERY exciting π π
lime trellis
haha post them in case other folks are interested. But I already hunted your fork down.
https://github.com/sajattack/atsamd
exotic pumice
lol, it should be merged soon into https://github.com/atsamd-rs/atsamd and gitter is https://gitter.im/atsamd-rs/community
marble hornet
woohoo
lime trellis
@exotic pumice would love to get a SAM32 into your hands to tinker on with atsamd-rs!
https://github.com/maholli/SAM32
exotic pumice
ESP32 would be totally new to me,but you could add a board def for the M4, which exact chip are you using?
lime trellis
The J20A
exotic pumice
ah ok, so it would be a tiny bit more work than a board def, we'd need to add the pac for that chip too.
lime trellis
ESP32 is actually most easily interacted with via ladyada's ESP32 SPI library
exotic pumice
@lime trellis, so that library would need a rust version too
lime trellis
It would be cool to get Phil Levis excited about this. He's the guy behind Tock http://csl.stanford.edu/~pal/
exotic pumice
cool
marble hornet
@meager fog thank you for the pyp tool! it is really useful
raven canopy
@marble hornet wrt ^^^, I haven't even read any of displayio, but are Shapes static or dynamic? If dynamic, zero could be allowed, but handled better?
tulip sleet
@slender iron i'm doing a big sweep of all the config files. framebuf should be off by default everywhere, right?
slender iron
@tulip sleet nice! yup
tulip sleet
yes, this is massive, but it should be a lot better in the long run, no more EXTRA_BUILTIN_MODULES and stuff like that
slender iron
great!
tulip sleet
also, if I conditionalize all the module defintions, maybe i can get rid of the WEAK_LINKS stuff? Do you know a strong reason to keep that? like the _time and other such modules, etc.?
this is relying on -flto to discard compiled but unused modules
i mean this refactoring is relying on that
I could put the on/off definitions in the .mk files, but right now they are in the .h files, and we rely on -flto to discard
slender iron
its meant to allow default implementations
slender iron
I know the interrupt handlers use weak
be careful that you aren't getting yourself in too deep
this reorg will be good but its a large task
tulip sleet
yah, but that's not MICROPY_PY_MODULE_WEAK_LINKS; that's what I mean I'm getting rid of, not weak links in general
tulip sleet
yes, it seems to be leftovers that we've moved to shared-bindings, mostly
slender iron
ya, I'd believe that
tulip sleet
ok, the proof is in the compilation, which will take a while to check everything
hopefully I can get -flto working on nrf too, though that's not vital now since we have the space
@slender iron displayio should be usable on nrf, right? Just checking for root pointer definitions
slender iron
yup!
tulip sleet
great!
slender iron
I don't think I added parallel bus but the fourwire should work
manic glacierBOT
rare geode
Owner of Python Discord here! Really looking forward to getting into circuit python, going to order myself some kit this week and have a play around, it looks awesome
rare geode
Oh wow, looking through this page https://learn.adafruit.com/sensor-plotting-with-mu-and-circuitpython?view=all, I need to get my hands on a Circuit Playground Express
that Mu integration is fantastic as well, kudos to whoever is behind that!
idle owl
@rare geode Group effort, but @plucky flint is the primary author of Mu π
rare geode
It's awesome! Great idea to use print statements and tuples, I love it
raven canopy
just one left...and its stubborn!
vagrant@ubuntu-xenial:~/source/circuitpython/lib/tinyusb/src/osal$ git stash
No local changes to save
vagrant@ubuntu-xenial:~/source/circuitpython/lib/tinyusb/src/osal$ cd ../../../..
vagrant@ubuntu-xenial:~/source/circuitpython$ git submodule update --init --recursive
error: The following untracked working tree files would be overwritten by checkout:
src/osal/osal_freertos.h
Please move or remove them before you can switch branches.
Aborting
Unable to checkout '29b49199beb8e9b5fead83e5cd36105f8746f1d7' in submodule path 'lib/tinyusb'
hah! just keep throwing git commands until...i win!
vagrant@ubuntu-xenial:~/source/circuitpython/lib/tinyusb$ git branch
* (HEAD detached at 6d96b12)
master
vagrant@ubuntu-xenial:~/source/circuitpython/lib/tinyusb$ git checkout master
warning: unable to rmdir hw/mcu/nordic/nrfx: Directory not empty
Previous HEAD position was 6d96b12... improve auto descriptor
Switched to branch 'master'
Your branch is behind 'origin/master' by 742 commits, and can be fast-forwarded.
(use "git pull" to update your local branch)
vagrant@ubuntu-xenial:~/source/circuitpython/lib/tinyusb$ cd ../..
vagrant@ubuntu-xenial:~/source/circuitpython$ git submodule update --init --recursive
Submodule path 'lib/tinyusb': checked out '29b49199beb8e9b5fead83e5cd36105f8746f1d7'
Try git submodule sync
raven canopy
i did. several times. to include:
git submodule foreach --recursive 'git fetch'
git submodule foreach --recursive 'git stash'
i had about 4 that complained before all of that, but tinyusb was so detached it needed some brute force. π
solar whale
I use git submodule update --init --recursive git submodule foreach --recursive 'git fetch --tags'
And I stand on my left foot π
raven canopy
ahh! that's probably it. i was sitting down. π€¦
main meteor
Is this expected when building CP? ```
../../shared-bindings/usb_hid/init.h:33:23: warning: size of 'common_hal_usb_hid_devices' differ from the size of original declaration [-Wlto-type-mismatch]
extern mp_obj_tuple_t common_hal_usb_hid_devices;
i think that is a known warning message. i've been building out of date, so haven't come across it.
main meteor
It's a really strange one, I wouldn't expect the size of a typedef to change, trying to wrap my head around how that could even happen.
Ah, it looks like it's a conflict between an array-in-place and an array pointer. Tricky.
tulip sleet
@main meteor are you using a version of arm-gcc older than 7.3.1? That looks like something fixed in 7.3.1.
@slender iron exciting build results: we've been relying on -flto to excise unused code, but it's still leaving remants. If I actually excise the source files in the Makefile conditionally, we get back hundreds of bytes from modules that aren't used
tough flax
Back from the frozen northwest. Thanks everyone for the HID help. Anyone happen to know if a ps4 will take an HID game pad? Or if thereβs a breakout board that emulates a ps4 controller?
manic glacierBOT
main meteor
Sure enough, I clicked on the wrong one when I backgraded from gcc 8 to 7. I grabbed the q4 one instead of q2, missing the detail that it was q4 2017. π³
tulip sleet
@slender iron following up on build results after compiling only source files that are really needed: trinket m0 build has 1936 more free bytes after doing this. Will try more builds. It's all working out nicely. Had to move a lot of setting stuff from .h to .mk, but it's all pretty clean. Time to π€
plucky flint
@idle owl @rare geode these things are always a group effort.... the "lone programmer" is a myth. We all depend so much on the work of others, which is one of the main reasons community is so important. π
slender iron
@tulip sleet That's epic! I think french is our largest language so try that too!
pastel panther
how does one get rid of a code.py that has their board in a boot loop?
remembers the debugger in front of him
hmm... no dice
(β―Β°β‘Β°οΌβ―οΈ΅ β»ββ»
π€
timber mango
@pastel panther yes, I would love to see more of your "itsybitsy -knockoff", do you have a link or github-repo ?
solar whale
@pastel panther boot to safe mode. Slow double tap, about 1 second between taps. Then you can get to REPL and import os then os.remove(βcode.pyβ)
tulip sleet
@slender iron on Metro M0, looks like the fr build has 1396 bytes free now as opposed to 244.
note that metro m0 has wiznet ethernet support turned on by default (that's not new)
the more features included in a build, the less you save (since more of the files would have been compiled in anyway)
manic glacierBOT
marble hornet
WOOHOO gui 1.x is done ! 2.x will include dislayio support.
manic glacierBOT
Ozone screensho after trap to HardDault_Handler -- unsupported return stack adderss -- not much information
manic glacierBOT
pastel panther
@solar whale No dice; tried various speeds of double taps, the best I can do is get into bootloader mode; I tried using the flash eraser from the troubleshooting page but macos says it's 141k too large and won't copy it π
solar whale
what version of CP did you have loaded and on what board
pastel panther
a fork off master (working on bitmap stuff)? I may have also flashed whatever the latest release is
solar whale
what board?
pastel panther
Grand Central
solar whale
just a sec -- I can make you an eraser -- maybe
been awhile -- trying to fin the correct spot.
tulip sleet
init fileystem in main.c -- make last arg true, or something like tthat
solar whale
try this
solar whale
diff --git a/main.c b/main.c
index c1244f36c..e81392027 100755
--- a/main.c
+++ b/main.c
@@ -389,7 +389,7 @@ int __attribute__((used)) main(void) {
// Create a new filesystem only if we're not in a safe mode.
// A power brownout here could make it appear as if there's
// no SPI flash filesystem, and we might erase the existing one.
- filesystem_init(safe_mode == NO_SAFE_MODE, false);
+ filesystem_init(safe_mode == NO_SAFE_MODE, true);
// Reset everything and prep MicroPython to run boot.py.
reset_port();
definitely filing that one away in my bag of tricks
@tulip sleet any idea why the GC flash eraser build on the troubleshooting page would show up as being too large but a freshly built one would not?
tulip sleet
are you sure you were trying to copy it to the boot drive and not the CIRCUITPY drive?
otherwise I don't know, but that's an arduino program, and could have been built wrong
(I have made the copy mistake in the past)
pastel panther
ya; I couldn't get a CIRCUITPY drive because of the boot loop so I'm sure it was the boot drive
tulip sleet
what's the name of the uf2 you were using?
pastel panther
GC_M4_QSPI_Erase.UF2
tulip sleet
i do not know - I'l try it myself later. afk for a coupla hours now
pastel panther
ttyl
thanks for the help
yay, still-in-bed-half-asleep-hacking is the best; 565 support here we come!
slender iron
@pastel panther update your repo. I fixed safe mode so it shouldn
't boot loop anymore
pastel panther
ok, thanks
slender iron
bootloader was overwriting the safe mode state
pastel panther
ah
umbral dagger
Is there a schematic around for the PyPortal?
slender iron
you have eagle? what form is best?
pastel panther
are there spare PyPortals around? π€£
umbral dagger
eagle is good. PDF is fine. I just want some details.
slender iron
@pastel panther not yet. I have an old rev myself
pastel panther
greens overflowing... π€
prime flower
@gilded cradle I have a pyportal on my desk if you'd want me to review the WiFiManager PR?
idle owl
@slender iron When you get a moment, can you look at this? I know you had an idea in your head of how to fix this issue and I want to make sure it's in line with that. https://github.com/adafruit/Adafruit_CircuitPython_Register/pull/21
gilded cradle
Sure @prime flower
slender iron
@idle owl sure, looking now
idle owl
Thanks!
gilded cradle
I've only tested it with a MetroM4 + Huzzah32 Feather
idle owl
@slender iron Thank you!
slender iron
np
marble hornet
@slender iron is mp_obj_t* children; the list inside of groups that holds the items ?
idle owl
@gilded cradle bump is the word Limor uses for creating a new release. Because you're "bumping up" the release tag number. π
gilded cradle
Yeah, just trying to learn some of the terminology.
idle owl
I understand entirely! I figured I'd make it even clearer if I could.
gilded cradle
Yeah. When you said I was already bumping, I was pretty sure what that meant. Just wanted to verify.
marble hornet
@slender iron i can't inherit from two objects that each individually inherit from Group
File "main.py", line 3, in <module>
File "/lib/tg_gui/gui.py", line 761, in <module>
TypeError: multiple bases have instance lay-out conflict```why would you inherit from two things?
marble hornet
i have a rect and a pointable object and i want to make a button
prime flower
@gilded cradle just tested it on my pyportal, it's awesome.
marble hornet
text is added ontop
prime flower
I like the ability to include headers as a dict
slender iron
@marble hornet I'd have button inherit from Group and then have rect and pointable object as children
prime flower
@gilded cradle Do you think adding in code for handling disconnects from a router (AP drops off) would be useful?
gilded cradle
Yeah, that's why it checks to see if it's connected on each call
prime flower
self.connect()``` Oh, it pings on each `get` or `post`Yeah
prime flower
good idea
gilded cradle
Thanks. That was part of the original request.
prime flower
Adafruit IO Arduino does a ping at the top of each loop to stay connected to the api/broker, this is way more elegant.
gilded cradle
π
prime flower
I'm going to try to whip together a really basic Adafruit IO wrapper to sit on top of WiFiManager
marble hornet
@slender iron what about inheriting from both causes the conflict?
marble hornet
Is this the double diamond problem?
slender iron
never heard of that
marble hornet
slender iron
my issue is that you rarely want a combination of APIs
more often you want a smaller api that wraps the two
and simplifies how to use them
marble hornet
can you say more about that? i'm not picturing how
ancient sage
What's the most common adafruit part used on/with a circuitpython board?
idle owl
@ancient sage That's a tough one. There's support for a lot of things. Maybe NeoPixels because who doesn't love LED rainbows?
The Circuit Playground Express has a ton of stuff built into the board, so it gives you a lot of options for learning different things in one package.
ancient sage
I'm going to try to support people with CircuitPython and I'm just wondering what would give the broadest range of testing. Already going to get a C.P.E.
idle owl
I think understanding the protocols, e.g. I2C and SPI, etc, is a good way to start. Because while different sensors do different things, if they're I2C and you know how to troubleshoot that, you're covered for a ton of different sensors. If that makes sense.
Check out the CircuitPython Essentials guide if you haven't already. It covers the different things built into CircuitPython, which most of our sensors/breakouts/etc use.
idle owl
Nice!
slender iron
@marble hornet have button subclass group and then in __init__ create the rectangle and add it to self's children (since self is a group)
prime flower
@gilded cradle io.send() workin'. I'm going to toss this in a repo before I break for lunch
need to figure out a better way of constructing the URLs passed into the $path though
gilded cradle
Awesome work @prime flower!
marble hornet
thank you, i think i have some ideas
exotic pumice
mac people, is there some package manager that has arm-none-eabi-gcc or do y'all just install the tarball every time?
slender iron
brew used to have it
exotic pumice
I couldn't find it searching their package list online
slender iron
they removed it because it was a cask and shouldn't have been π
main meteor
I just install the tarball. I tried using the gcc that comes with Arduino, but it's not the right vintage.
slender iron
exotic pumice
speaking of vintage compilers, I wrote some 6502 assembly for C64 yesterday
it just endlessly prints "green on black" in green on black
main meteor
Nice! Tempted to fire up my old 6800 machine and see if it still goes.
pastel panther
@exotic pumice the brew cask people removed it because they said it belongs in regular brew but they didn't think to/want to /have time to add a PR to add it to brew
I was planning on submitting a PR but haven't had time
exotic pumice
got it, thanks
slender iron
I just asked to see if someone is working on it
exotic pumice
is cask binary package versus source package or something?
slender iron
ya
pastel panther
their github page also mentions targeting GUI apps
solar whale
@slender iron If I want to increase the stack for the nrf52840 build do I only have to change the setting in mpconfigport.h -- or are there other places as well?
slender iron
I think just there
solar whale
thnaks -- I'll try increasing it to see if it helps
slender iron
it'd be good to add stack checks so this fails more gracefully next time
solar whale
is that an mpconfigport.h setting?
yup -- doubling the stack got rid of the problem ....
solar whale
at least for the simple test -- trin ghte full test next
solar whale
so in does atmle report stack when going to safe mode, but nrf does not -- it issed HardFault
manic glacierBOT
slender iron
right, we can only go into safe mode if we check before we crash
gilded cradle
I think it is going into safe mode
solar whale
right -- it goes to safe mode but says it was due to HardFault, not stack overrun.
slender iron
ideally we'd notice we're out of stack and throw an exception before we crash
solar whale
it did taht for me in another test, so it may depend on how it runs out of stack
tulip sleet
stack checking was turned off in the nrf port, for some reason. I noticed that while refactoring. It will be on after my big PR.
also the default stack will be 24k
slender iron
ah, so it can be turned on?
tulip sleet
MICROPY_STACK_CHECK
solar whale
great! issue solved π
slender iron
may be worth trying jerry
gilded cradle
π
slender iron
just to see if it fails nicer
solar whale
OK will try
gilded cradle
We should try on the Argon too
tulip sleet
argon is same
no stack check and 4k check are common across all the nrf boards
as of now
solar whale
gdb was not much help!
slender iron
in gdb I like to break on reset_into_safe_mode
solar whale
I had break on HardFault_Handler, and it worked ok -- more my lack of familiarity with what was reported -- rusty on gdb...
Even with stack checking enabled, it went to HardFault
slender iron
hrm
hardfault_handler is good too
it'll call reset_into_safe_mode too
the checks may not be in the right places
solar whale
I posted the backtrace to the issue -- you can read it at your leisure -- it may tell you more than it did me.... I guess I'll leave the issue open until Dans PR is in -- I'll add some notes
umbral dagger
I just wrote a better repl (read more than readline from stdin) for CircuitScheme. Basic readline move left/right, backspace/forward-delete, history using up/down, multi-line command input.
Would a guide be of interest? It's general purpose (just a parse and eval ties it to CircuitScheme).
It's pure CircuitPython code.
manic glacierBOT
solar whale
@gilded cradle argon works!
gilded cradle
Sweet
solar whale
Whew! I don't have many hairs left -- lost quite a few on that one π
gilded cradle
I got an Argon too. I just haven't put the bootloader on it yet.
I mean it just got here today.
idle owl
Tape 'em back on, @solar whale!
solar whale
@gilded cradle ave fun with the Argon -- If you put the AT Firmware on the ESP32 you might get inspired to add the Wifi Wrapper to the ESPATControl library! Let me know if you have any problems getting it going.
gilded cradle
Ah, ok. I might try that.
exotic pumice
Twitter
solar whale
@exotic pumice this was a great exercise for me -- I need be quicker to pull out the J-Link and gdb. It's a thrill!
exotic pumice
hahaha
ruby tiger
how much storage space would typically be used by circuitpython's code and libraries? assuming this is a small project with not much code. rough estimate is fine.
river quest
@ruby tiger it's small, so even with images and sounds, everything does not fill up the "USB" drive
ruby tiger
but like, on the playground express, how much of those 2MB can I actually use to store media?
prime flower
@gilded cradle Can I extend your WiFiManager for DELETE/PUT?
gilded cradle
Sure @prime flower, that works for me.
prime flower
They're within requests.py, not sure if you are going to add them or not and dont want to overrun your work
river quest
@ruby tiger almost the full 2mbs
ruby tiger
super cool. thanks @river quest
prime flower
ah np, they work great
super early and ugly (not for public consumption unless you really want to break things) , but ya can pass it in the wifi manager
gilded cradle
Nice. That should make doing stuff with Adafruit IO even easier.
ancient sage
Can I do THUMB-2 assembler inline in CircuitPython on the ATSAMD21?
exotic pumice
@lime trellis I'm writing docs on adding new chips and boards to atsamd-rs https://github.com/atsamd-rs/atsamd/wiki/Adding-a-new-chip
ancient sage
Being a strict subset of cpython is a noble goal as well π
I was a little freaked out by the form of the files in tests/ from micropython at first, but it seems that the strategy is to run everything against cpython and micro/circuitpython and check for differences?
it was like, these are the worst tests I've ever seen
tulip sleet
@ancient sage the tests do catch errors. If you think those are bad, wait until you see the ones we don't have.
ancient sage
Haha... Is there a need for more tests?
Initially I thought that it only ran the test files and checked for exceptions
Thus my astonishment
tulip sleet
we need more hardware-based regression testing. Those tests were written by the micropython folks. We have the start of some hw tests, but it's not automated yet. We had an automated tester that ran the test suite on several actual hardware boards, but it was kind of flaky and we haven't revived it (it lives at tannewt's house). https://github.com/adafruit/Adafruit_CircuitPython_BoardTest
that's the board test library. The automated tester is called rosie, and didn't do pin checking. the code for that is elsewhere
the micropython based tests of course are run on travis builds
ancient sage
Right, that wouldn't catch anything that only shows up when cross compiled
tulip sleet
the mpy tests are quite good at catching regressions in the core language impl, which we don't mess with much
@slender iron do you have a strategy for dealing with translation merge conflicts? Maybe I should just do git checkout --theirs ... and take the upstream one?
slender iron
@tulip sleet I think I take theirs and then rerun make translate
tulip sleet
that's what I think, but I need to make sure it's doing the right thing. ok tnx
aha, needed to do git add after git checkout --theirs kk
raven canopy
@tulip sleet @idle owl for the bundle link description, i'm thinking of a merge of both ideas. something like "Documentation for the bundle, which includes links to documentation for all libraries". thoughts?
idle owl
@raven canopy Sounds good
tulip sleet
that sounds good -- i didn't really know what it was actually pointing to
is there a summary page of all the libraries you could link to instead of just the bundle RTD link?
instead of or in addition to?
raven canopy
kk. commit imminent...or wait. π
idle owl
That's what the bundle RTD is
tulip sleet
shows how much i know
raven canopy
yeah, we just moved the "old" drivers.rst list to the bundle.
raven canopy
π
idle owl
Bundle RTD is links to all the libraries. And being added to the process of adding a lib to the bundle, is updating that file.
That way we're not forking CP just to update that drivers file anymore.
And it'll be kept up to date!
tulip sleet
i love it
raven canopy
and now...we wait for Travis to test two lines of text. π°
exotic pumice
Anyone wanna add a board to atsamd-rs? π https://github.com/atsamd-rs/atsamd/wiki/Adding-a-new-board
raven canopy
@exotic pumice is knowing Rust a requirement? π
idle owl
@raven canopy So.... idea....
exotic pumice
@raven canopy barely
having it installed and following my instructions should be enough lol
raven canopy
uh-oh...ideas.
idle owl
@raven canopy In the bundle RTD under "Adding a library" should we include instructions for updating the docs page? I mean we need to update the guide on "Creating and Sharing a Library", but should it also be in the Bundle RTD?
Also my weirdo caching issue is magically fixed so that's nice.
raven canopy
that was the thing i kind of noted on the PR. i almost started to put in a "how to contribute to the bundle" section, linking to the guide and having an outline.
idle owl
Ahhh ok
I remember that in the PR, wasn't thinking it was the same thing. Obviously it is.
raven canopy
i mean, it costs nothing (or, very little?) to add it.
idle owl
Hmm. Let's get the guide updated first maybe?
Mostly so you don't have to do more right now π
raven canopy
ahh, that's just a little creatively technical writing. always cycles for that. π after deciding dinner and retrieving it of course. hehe
idle owl
You want to update the guide too? π
raven canopy
walked into that one. π€¦ i mean, of course i do!
idle owl
π
raven canopy
seriously though, yeah, i can add it. though i think i'll need access to it.
idle owl
Nah that wasn't yelling. It was meant to be more of a creepy whisper.
But well played π
exotic pumice
π
solar basin
Woohoo! My Adafruit box came in today!
I have goodies for days!
I didn't realize that the JTAG ribbon cable was so small
tulip sleet
pins are lot smalle than 0.1" spacing
manic glacierBOT
gilded cradle
Oh right, the Adabox. Mine hasn't shipped yet and I'm about to move, so I just changed my address to my friend's house since it might come after I move.
solar basin
I just bought an Adabox 10 ad-hoc
gilded cradle
Ah, that makes sense. I think the next ones ship next month.
solar basin
That is my understanding as well
gilded cradle
But just in case I forgot to change my addr, I set it to something where I could still get it.
solar basin
Nice!
tulip sleet
yay
pastel panther
why do the pixels hate me so? π’
main meteor
Fighting displayio?
pastel panther
kinda; trying to add 565 support to OnDiskBitmap and my green channel is overflowing
main meteor
I was just looking over OnDiskBitmap last night, amusingly enough.
pastel panther
I thought I had my bitmask wrong but the input pixel from the disk seems to agree with what I'm getting out so either what's coming in is wrong or my expectations are
tulip sleet
@slender iron the last several travis merge builds have failed, not sure why. Restarting doesn't help. The shallow clone claims something has changed. This happened before my big PR. The PR builds are fine. π€·
pastel panther
signs are pointing to me being wrong....
Success!
Don't over shift your bitz, boyz and gurlz
ancient sage
pastel panther
A for effort?
tulip sleet
@raven canopy are you happy with the Move the Docs PR? I can merge it. All the merge builds seem to be failing due to depth 1 and a lot of outstanding merges, I think.
raven canopy
hehe. i just kicked the tires on the Travis for that PR. even after reading your post. i'm good with the PR. want me to stop the Travis?
tulip sleet
no problem. I'll just wait and see, or approve anyway. I'm not sure why the depth:1 is fialing but it might be due to so many simultaneous PR's.
raven canopy
it seems to be past the funky bits.
tulip sleet
the PR builds are fine, it's the merge builds after the merge is doen
raven canopy
ahh. my last PR run failed with the same error, so i thought it was across the board.
tulip sleet
ok, didn't see that, but I'll wait for it to finish before I merge
raven canopy
π
tulip sleet
https://stackoverflow.com/questions/31278233/disadvantages-of-shallow-cloning-on-travis-and-other-ci-services but I'm not sure some of the answers are correct (@slender iron take note too)
raven canopy
hmm. interesting theories...
tulip sleet
the accepted answer does seem to be wrong about the commit hash. but we have a backlong of multiple commits and a depth of 1 wouldn't fetch the right one if there were merge commits of other PR's in between.
i will knock off a bunch of these and re-run the top one if necessary
gilded cradle
I just got CircuitPython running on my Particle Argon. I knew buying a JLink would come in handy down the road. π
lol
raven canopy
@tulip sleet @slender iron from the Travis docs on clone depth:
Please note that if you use a depth of 1 and have a queue of jobs, Travis CI wonβt build commits that are in the queue when you push a new commit.
https://docs.travis-ci.com/user/customizing-the-build#git-clone-depth
tulip sleet
that is definitive!
we have depth: 1 but we also have:
- git fetch --shallow-exclude=4.0.0-alpha.1 || git fetch --unshallow
- git fetch --depth 1 origin $LAST_TAG:$LAST_TAG
so I'm confused
ruby tiger
hey guys! does CPy support f-strings?
tulip sleet
@ruby tiger no f-strings - it's not in MicroPython either
ruby tiger
aww. that's too bad, they're lovely.
raven canopy
@tulip sleet how's Travis looking this morning? still a bunch of βs?
he asks lazily, while waiting for coffee to work π
tulip sleet
I approved a PR a few minutes ago and it messed up too. If Scott is not around this weekend, we can discuss it on Monday/Tuesday, about increasing the depth, etc. I need to go back and look at the old changes to see why we did what we did.
raven canopy
i would hazard a guess that it was done in the interest of speed..
tulip sleet
I don't think the clone time is the major bottleneck now. It's the zillions of translations. Making a depth of 6 or something might be fine.
raven canopy
tulip sleet
The extra git fetch is supposed to pull more as necessary, but it doesn't seem to be working. I should try this by hand locally and see what's going on.
raven canopy
and the associated PR: https://github.com/adafruit/circuitpython/pull/1360
raven canopy
oof. re-gluing code in and trying to remember why i had done certain things. taking 4 month breaks on something has detrimental effects. π¬ π€·
main meteor
I know that feel.
raven canopy
eureka! now i remember. π
sinful thorn
hiya :) i seem to be unable to connect to the REPL on my circuit playground express. saving and modifying the circuitpython code works fine and it behaves as expected, but mu shows this error when attempting to open a serial connection. i've pressed the reset button but that doesn't help either. i've also had a look at this https://learn.adafruit.com/welcome-to-circuitpython/advanced-serial-console-on-mac-and-linux but nothing shows up when i do ls /dev/ttyACM* (i'm running arch linux, btw). is there perhaps anything else i could try? :D
raven canopy
@sinful thorn try lsusb. does that list the device?
sinful thorn
hmm, not from what i can see: ```
Bus 003 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 007 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
Bus 006 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 005 Device 003: ID 256c:006e
Bus 005 Device 022: ID 239a:8019
Bus 005 Device 002: ID 093a:2521 Pixart Imaging, Inc. Optical Mouse
Bus 005 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
Bus 001 Device 002: ID 058f:6366 Alcor Micro Corp. Multi Flash Reader
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 004 Device 002: ID 045e:07f8 Microsoft Corp. Wired Keyboard 600 (model 1576)
Bus 004 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
yep. Bus 005 Device 022: ID 239a:8019. now, does ls /dev/ttyUSB show anything?
raven canopy
is it sad when USB VID/PIDs are burned in your memory?
sinful thorn
not particularly :D
raven canopy
i'm trying to find it, but i seem to remember a conflict with modemmanager...
sinful thorn
well, i don't even have modemmanager installed.
raven canopy
sweet. cause the "fix" is to remove it. haha
sinful thorn
yeah i read about that when looking for a solution myself, but i didn't even have it installed in the first place so.. Β―_(γ)_/Β―
raven canopy
just for grins and giggles, what does ls /dev/tty* spit out?
sinful thorn
/dev/tty /dev/tty19 /dev/tty3 /dev/tty40 /dev/tty51 /dev/tty62
/dev/tty0 /dev/tty2 /dev/tty30 /dev/tty41 /dev/tty52 /dev/tty63
/dev/tty1 /dev/tty20 /dev/tty31 /dev/tty42 /dev/tty53 /dev/tty7
/dev/tty10 /dev/tty21 /dev/tty32 /dev/tty43 /dev/tty54 /dev/tty8
/dev/tty11 /dev/tty22 /dev/tty33 /dev/tty44 /dev/tty55 /dev/tty9
/dev/tty12 /dev/tty23 /dev/tty34 /dev/tty45 /dev/tty56 /dev/ttyS0
/dev/tty13 /dev/tty24 /dev/tty35 /dev/tty46 /dev/tty57 /dev/ttyS1
/dev/tty14 /dev/tty25 /dev/tty36 /dev/tty47 /dev/tty58 /dev/ttyS2
/dev/tty15 /dev/tty26 /dev/tty37 /dev/tty48 /dev/tty59 /dev/ttyS3
/dev/tty16 /dev/tty27 /dev/tty38 /dev/tty49 /dev/tty6
/dev/tty17 /dev/tty28 /dev/tty39 /dev/tty5 /dev/tty60
/dev/tty18 /dev/tty29 /dev/tty4 /dev/tty50 /dev/tty61
forgot to mention actually, i believe the /dev/ttyS* would be serial ports but i just got a blank screen when attempting to use screen to connect to /dev/ttyS0 (which is the only one which didn't immediately terminate with or without sudo) with a baud rate of 115200.
raven canopy
really just stalling to avoid sifting through dmesg π
sinful thorn
haha, i avoid my own dmesg when i can. :D
raven canopy
yeah, nothing there looks like it would be it. i have no arch experience, but kernel is kernel, right?
sinful thorn
how do you mean?
tulip sleet
if you are running a main.py, screen to /dev/ttyS0 will be blank until you type ctrl-C
raven canopy
@tulip sleet have you Arch experience? would it show up on ttyS*?
tulip sleet
no, no arch experience. what version of arch? Are you up to date?
do you have any other computers in the house, Windows or Mac?
sinful thorn
i did try pressing ctrl-C, didn't work either. i'm running kernel version 4.20.6 (i upgraded to 4.20.8 earlier but haven't rebooted since, actually). all of the computers in my house are running linux. i could dig around for an old windows laptop, but i think it still needs windows 7 SP1 haha
tulip sleet
all arch, or any debian or ubuntu
sinful thorn
there's a linux mint computer somewhere, which is ubuntu-based.
not mine though, so i'd have to wait until it's not in use.
tulip sleet
maybe I can find a DVD-boot arch to try, or find an external disk to boot from.
looks like you should see /dev/ttyACM0
sinful thorn
i don't, unfortunately. my entire ls /dev/tty* output is shown above
tulip sleet
what version of CPy are you running on the CPX?
sinful thorn
version 3.1.2
tulip sleet
ok, that's good - i see some arch linux live-cd/usb instructions. i'll give it a brief go. will get back to you later. in the meantime maybe bother the linux mint person if available
sinful thorn
will do. :D
raven canopy
@sinful thorn what type of port do you have it plugged into (USB2, USB3)? shouldn't matter, but long-shot
raven canopy
have any other ports available to try? other than that, its a wait on dan or dmesg. dan is much smarter than i...so i'll defer to him. π
sinful thorn
yeah i'll try a few others.
tulip sleet
flashing a usb drive...
sinful thorn
works perfectly on linux mint. :)
tulip sleet
I made a bootable arch 4.20.6 usb drive, booted it, plugged in a CPX, and I see /dev/ttyACM0. You said you updated but haven't rebooted yet? So maybe reboot and see if that fixes things.
sinful thorn
okay sure.
heck of course that works haha. reboots fix everything, don't they? sorry for wasting your time, thanks for the help :D
raven canopy
yay! π
tulip sleet
ok, wait, this is interesting. I pressed the reset button and then /dev/ttyACM0 did not reappear, or when I unplugged it and plugged it back in. Try unplugging and replugging and see if it appears
That was after doing cat </dev/ttyACM0
@sinful thorn ^
tulip sleet
@sinful thorn ok, good! I was getting weird errors, but quite possibly due to booting from USB drive
sinful thorn
that's possible, yeah. everything's working fine for me though, thanks :D
slender iron
<@&356864093652516868> Here is the notes doc for next week's meeting on TUESDAY! https://docs.google.com/document/d/1Nd29jwuRkcS4M415TvGDRGjx6aA2M3xzeYl5m-gPJaI/edit?usp=sharing
idle owl
@slender iron I put "Community News" after the State of CircuitPython in my template π I'll move it to before. Mostly I'm entertained that we chose different places for it.
pastel panther
@tulip sleet is it kosher to have PRs targeting master coming from a branch on the fork, or is it preferred to have them coming from master on the fork?
idle owl
@pastel panther If you're asking what I think you're asking, branches are fine. As long as it's up to date and you know it is, it shouldn't matter.
pastel panther
kk
tidal kiln
id say that's fine and generally preferred. i try to do it that way and make the branch name reflect the point of the PR, like give it an issue number, or feature name, etc.
and. well. there's this learn page with a pretty direct title:
https://learn.adafruit.com/contribute-to-circuitpython-with-git-and-github/always-work-on-a-branch
tulip sleet
@pastel panther I always use branches so I can keep my work organized and save a line of development away if necessary
cater and kattni are on the money
pastel panther
I do too! For some reason I'd previously always made PRs from master and I didn't know if that reason was a rule or something I just did arbitrarily
idle owl
Hey I wrote that guide! π
tidal kiln
i'm not skilled enough to have multiple branches going at one time though
raven canopy
@tidal kiln its a total juggling nightmare. i've gotten into that state only a couple times, but...ugh.
tulip sleet
the important thing is to make sure to commit before you switch branches. you can also use git stash, but that's harder to keep track of.
or just clone again
gilded cradle
Sometimes I work on multiple branches for the Adafruit_CircuitPython_FeatherWing library. I just merge them into master when the branches are ready to go.
tulip sleet
git clean -xf will clean out build junk, etc.
marble hornet
@slender iron I'll have to miss the meeting tomorrow, robotics. But I did make a pwm tool with circuit python to help with the robot. Best of luck to all and a large group hug
idle owl
@marble hornet The meeting is on Tuesday this week.
Also tomorrow is Sunday π
We're making the notes doc on Saturdays now to provide more time to enter notes into it.
raven canopy
@slender iron i only sort of understand what is going on with Travis, so I'm going to leave that ^^ PR to you for approval.
umbral dagger
@pastel panther I always use branches for new work as well.
slender iron
k, I'm gonna look next week. shouldn't need to rush it
exotic pumice
uh, how is dotstar_sck on pad 0 in the neotrellis m4?
my understanding was sck could only be pad 1 or pad 3
raven canopy
@exotic pumice i see I2S/SCK[0] in the I/O for PB16. pg 34.
raven canopy
can be either, is how i've understood that. how to choose... π€·
well, not either. but...you know what i mean. i hope. what are we talking about? lol
exotic pumice
dotstar_clk is pb02, which is sercom5pad0 which can't be sck as far as I can tell https://cdn-learn.adafruit.com/assets/assets/000/064/889/original/adafruit_products_mainschem.png?1541024554
raven canopy
well, nvm. seems SCK in that instance is related to I2S?
exotic pumice
yeah
raven canopy
you threw me off with the SCK. too used to SCLK or CLK. π
raven canopy
yeah, i'm tracking now. SERCOM5/PAD[0] is available for PB02
raven canopy
it uses bitbangio, not busio: https://github.com/adafruit/circuitpython/blob/master/ports/atmel-samd/boards/trellis_m4_express/mpconfigboard.h#L11
exotic pumice
ah darn
was that pin choice a mistake I wonder?
if you swap sck and mosi then you can use a sercom
raven canopy
may have been more to keep the sercoms available for other things?
exotic pumice
maybe
I don't have a rust bitbang apa102 driver, so I'll have to move onto something else
I find the "keep sercoms free" theory implausible, aren't you only using 1 sercom for i2c?
raven canopy
π€· it was just a thought. ladyada is way more smarter and experienced on board design than i'll prob ever be. π
gilded cradle
If I wanted to compile the latest CircuitPython on my local computer for a particular board, what command would I use? I think I might already have all of the dependencies (I've successfully compiled bootloaders and mpycross before).
Well, for now I'll just grab the latest from https://adafruit-circuit-python.s3.amazonaws.com/index.html?prefix=bin
tired steppe
Anyone here have experience building CP on linux?
solar whale
@gilded cradle this guide should help. https://learn.adafruit.com/building-circuitpython
woeful juniper
Hi there, I'm newbie. Actually I don't know difference between circuitpython and micropython. But I have a problem about micropython's re library. It have not DOTALL flag. How can I match (.) special charachter with newline characters too?
solar whale
@tired steppe yes, what information are you looking for?
@woeful juniper I don't think I understand your question - can you provide an example of what you are trying to do -- the newline can be represented as "\n" - carriage return as "\r" if that helps.
@gilded cradle the actual build is done much like the bootloader make BOARD=feather_m4_express clean then make BOARD=feather_m4_express for example. The tricky part is getting the full repo and tools installed but you may have most of that done. This page has the crucial stull https://learn.adafruit.com/building-circuitpython/build-circuitpython#fetch-the-code-to-build-6-1 and I would ad one thing after you clone or update the repo do git submodule update --init --recursive git submodule foreach --recursive 'git fetch --tags'
woeful juniper
@solar whale Thanks for attention, I'm just trying create simple template engine so I need to parse template string. I'm trying split with re string with {{(.*?)}} template and re.DOTALL flag. It's not matching {{thing\nsomething\n anotherthing}} without DOTALL flag. As you know the dot char (.) matching everthing except newline. But this flag is not available on micropython. How can I handle it?
solar whale
@woeful juniper sorry - you are well beyond my knowledge of regular expressions - Hopefully someone else can help later in the day. Good luck!
woeful juniper
@solar whale thank you so much. Actually I have another idea its all about reading string char-by-char and parse it. But I'm not sure is it best way for performance. Do you have any suggestion about it?
solar whale
@woeful juniper just to be sure -- have you looked at the ure module http://docs.micropython.org/en/latest/library/ure.html
as to parsing each character, I'm not sure how much difference it really makes. I guess it depends on how often you are executing this code.
also this may help https://stackoverflow.com/questions/29020866/can-i-match-multiline-string-in-python-without-using-re-dotall
woeful juniper
@solar whale thanks again, I didn't see this topic on stackoverf. It looks like great solution.
solar whale
@woeful juniper great! You are welcome. Good luck!
tired steppe
@solar whale Nevermind, I figured out my problem last night and was able to build CP without any issues
manic glacierBOT
umbral dagger
CP keyboard handling head scratcher: I'm working on a REPL for CircuitScheme. It's coming along well (maybe done for now) except for supporting C-c. I'm catching KeyboardInterrupt and am getting it when C-c is pressed... but not until another key is pressed. I'm using sys.stdin.read(1) to get keys. Thoughts?
solar whale
@umbral dagger it works normally on your board in CP, correct -- or does sys.stdin.read(1) block it?
oh -- I see, it does
same for input()
umbral dagger
Fine in CP
solar whale
at least on the nrf build it does
umbral dagger
Having async C-c is pretty important. For the same reasons it is in CP
solar whale
for me in cp if so sys.stdin.read(1) or input() it does not respond to control-c until next char
umbral dagger
But I don't want it to break out to the CP REPL. Just the CS REPL.
solar whale
actually until next EOL
umbral dagger
I'm seeing it happen on next char
solar whale
clarify: input() waits for EOL sys.stdin.read(1) waits for another char
that is in CP
solar whale
right - now I understand -- looks like it blocks until at least one char is received, but the contol-C is not treated as a received char.
yikes -- don't do sys.stdin.read() w/o argument -- it just hangs
umbral dagger
Right. I never read the \x03 character
Which is fine since I need it to be asynchronous so as to interrupt running code
solar whale
still trying to find the code for sys.stdin
solar whale
doesn't look good π
gilded cradle
Thanks for the information @solar whale
Ok, yeah it compiled successfully without having to install anything new beyond submodules. Also the 2 time functions you mentioned to be missing are now gone on mine too, but the rest of it works.
manic glacierBOT
umbral dagger
@tulip sleet I built the latest for the nrf52840 feather. Now it's saying there's no io module. Shall I open an issue for that?
tulip sleet
Yes. Just 'io`?
umbral dagger
The rest are just the usual suspects.
So that should be it.
How do I add it? You've revamped that stuff, right?
tulip sleet
Whatβs the specific error?
umbral dagger
@tulip sleet ```Adafruit CircuitPython 4.0.0-beta.2-74-g90bc09a31 on 2019-02-17; Adafruit Feather nRF52840 Express with nRF52840
soft reboot
Auto-reload is on. Simply save files over USB to run them or enter REPL to disable.
code.py output:
Traceback (most recent call last):
File "code.py", line 31, in <module>
ImportError: no module named 'io'
Line 31: from io import StringIO
tulip sleet
@umbral dagger did this code used to work? Could you point me to the code?
umbral dagger
Works on SAMD51 boards
tulip sleet
did you ever test it on nrf before today?
umbral dagger
Yes, that's how I discovered that ure wasn't included.
I'm in the process of trying to get it running on the nRF feather (and eventually others)
tulip sleet
yes, but did you do a successful from io on nrf before today? I know it needs to be defined; I'm just trying to figure out if this is an old or a new problem.
solar whale
@tulip sleet ran into this -is is it expected when setting an int with a large value? ```Press any key to enter the REPL. Use CTRL-D to reload.
Adafruit CircuitPython 4.0.0-beta.2-72-gf7a397d52 on 2019-02-16; Particle Argon with nRF52840
import featherwing_rtc_simpletest
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
OverflowError: long int not supported in this build
this is the cause rtc.unixtime = 1550335257
tulip sleet
could you try on f840? Has something to do with size of build.
solar whale
ok -- is argon differnt from nrf52840?
tulip sleet
@umbral dagger @gilded cradle @solar whale I have a lead on missing time stuff, etc.
umbral dagger
@tulip sleet Have not imported io on an 840 board before. This is my first real attempt at CP on 840
tulip sleet
from io didn't actually work on f840 before, but it should, and I'm fixing it.
does argon have a spi flash chip/
solar whale
yes 8Mbytes
tulip sleet
ok, I'll check the definitions, might be an ordering problem or something
ok, tried on f840, longint is off, by mistake
will fix that too
solar whale
sorry only 4Mbytes
can you fix this too - argon does not mount the file system if you load the feather_nrf52840_express UF2 to it π€¦
I was almost ready to create an issue...
gilded cradle
Thanks @tulip sleet. Also, I'm not sure if this was intentional or not, but the Argon's debug LED does not light up for me (at least on the latest Alpha 2 release).
solar whale
Does CP even try to use the Argon LED? -- it is not a neopixel but an RGB LED
gilded cradle
Ohhh, that explains it. I think there's an issue to add an RGB LED to CP.
solar whale
and D7 blinks a blue LED (not D13)
tulip sleet
i missed the earlier discussion: which boards are missing ure in the latest master?
gilded cradle
What is ure?
tulip sleet
micropython's version of re. We should rename it to re sooner or later.
gilded cradle
Ah, ok.
manic glacierBOT
solar whale
sorry - I was not part of the ure discussion -- not sure where the issue was.
umbral dagger
I found it missing on the feather nrf52840 build.
And if we're revisiting ure/re it'd be nice to flesh it out.
It's missing character classes when I looked.
gilded cradle
It's missing on argon too
tulip sleet
ok, thanks, on the trck of that too
solar whale
@tulip sleet o good deed goes unpunished π
tulip sleet
ok, i have fixes for missing time.mktime() and missing ure. still working on io. It will all be in one PR.
i expected worse!
solar whale
will it also fix time.localtime?
tulip sleet
and longint fixed
is it that localtime() is missing altogether or isn't taking the right args?
solar whale
Adafruit CircuitPython 4.0.0-beta.2-72-gf7a397d52 on 2019-02-16; Particle Argon with nRF52840
>>> import time
>>> dir(time)
['__class__', '__name__', 'monotonic', 'sleep', 'struct_time']
>>>
ok, there is no rtc on nrf yet, so right now `time.localtime() throws RuntimeError: RTC is not supported on this board. should it do something else?
gilded cradle
it's missing altogether from what I compiled this morning.
solar whale
should it work with the RTC Featherwing ?
tulip sleet
CPy doesn't know if the featherwing is there or not, so I don't think so
solar whale
so only time.locatime needs the builtin rtc? everything else works for rtc via the library
tulip sleet
i guess, not sure what it does on boards without rtc enabled (or did do)
solar whale
I see -- it is time.localtime,, not rtc.localtime -- nevermind
solar whale
ok -- but for now all but the "unixtime" commands work on the rtc featherwing on the nrfs so that is good -- and time.locatime gives an message along wit the error.
gilded cradle
Well, that's on the latest master. It's been about 2 weeks since the latest official release of CP and that one chokes.
It chokes on the DS3231 simpletest too
solar whale
and there is no rtc.localtime -- the rtc.unixtime functions just call time.localtime
gilded cradle
I think rtc.localtime() is a coincidentally named low level function.
granite crow
Hi, i just got a feather nrf52840 board and i'm playing with it, this might be super easy mistake i'm doing but the pinout says the user switch can be used as D7 but i get the following error when trying to use it switch = DigitalInOut(board.D7) Traceback (most recent call last): File "<stdin>", line 1, in <module> AttributeError: 'module' object has no attribute 'D7'
idle owl
@granite crow From the REPL: do import board and then do dir(board). It might be something like board.SWITCH.
I don't have one connected at the moment to check.
tulip sleet
just looked at the source, yes, board.SWITCH
idle owl
There you go π
tulip sleet
says "D7" in the pinouts stuff User/DFU Switch
A tactile switch is provided for use in your projects, which is connected to P1.02 and is accessible in code as D7. Maybe we should say "Arduino pin 7 or CircuitPython board.SWITCH" or something like that
granite crow
great, it works! thanks for the fast reply
yep, also maybe with the user led can be used as led = DigitalInOut(board.RED_LED)
exotic pumice
what does the magic number 500000 represent in bitbangio spi?
tulip sleet
self->us_delay = 500000 / frequency; frequency is in Hertz. us_delay is microseconds delay, so this is half of a clock cycle
raven canopy
the lucky winner? π
exotic pumice
oh I see, 1Hz is 1000000us
tulip sleet
yah
exotic pumice
got it
tulip sleet
i agree a comment might be good
obsidian dome
can someone point me at any directions for how to build CircuitPython from sources?
gilded cradle
Yeah, I just had that question answered earlier today: https://learn.adafruit.com/building-circuitpython/build-circuitpython
obsidian dome
Thanks!
exotic pumice
what's the max speed of bitbangio spi?
my reimplementation is going 30KHz when I tell it to go 3MHz and I'm not sure if that's a math bug or a limit
seems like a limit
turning on optimizations got me to 150KHz, interesting
obsidian dome
Can anyone point me at what circuitpython components are excluded for the trinket_m0?