I hadn't read that
#circuitpython-dev
slender iron
raven canopy
I briefly did, when I noticed the register in Atmel Studio. Didn't even see it in the sheet before that ๐
I'm going to be out-of-pocket starting in a little while, and all day tomorrow. I'm shooting to have the PR in on Saturday. I was really hoping to go full on TC capture with the M4, but it just won't give up the ghost. ๐ฆ
slender iron
did you try my suggestion yet?
raven canopy
i haven't. i can give it a go Saturday morning/this weekend. so I understand fully, you're suggesting a stripped down PulseIn, yes?
instead of counting the width, just count the rises?
i also need to stop saying "tempted to DMA", and actually try it. ๐
slender iron
not exactly because pulsein uses an interrupt every edge
I'm suggesting using the TC to count pulses
(TC.EVACT = COUNT)
raven canopy
so use the COUNT, instead of it's internal... yeah.
slender iron
then you set the compare channel to the number of pulses you want to average over to generate the interrupt
essentially dividing down the input signal
raven canopy
hmm. i'll have to re-read to see how to interface the incoming edges with COUNT. I was under the impression that compare was only an out operation. it is totally possible my comprehension is off/biased. ๐
raven canopy
yeah, the overflow i still need to figure out how to deal with anyway. on capture, it can be used "calibrate" in the case of the captures happening after the count passes MAX. it'll be similar to how PulseIn handles the SysTick reload, I think. still have the goal of keeping overhead at a minimum to ensure the highest possible capture capacity without locking into an endless loop.
slender iron
right, thats why I think dividing down is the way to go
raven canopy
i can definitely attempt it. alright i'm off. lunch and the punishment of golf awaits us... ๐ ๐
slender iron
byeee!
errant grail
CPy MIDI sniffer mod for OLED wing. Work in progress. https://youtu.be/U_K38mXqN60
idle owl
Sniffing MIDI data?
errant grail
Yes. The version that sends output to the REPL or terminal window works nicely, but the OLED wing version occasionally blocks incoming data.
idle owl
Interesting. That's super neat though!
errant grail
Needed it to learn more about the MIDI protocol stack. I now know enough to be musically dangerous.
idle owl
An excellent thing to be
@tulip sleet I'm trying to follow the same process we're asking of everyone else. Can you approve this for me? https://github.com/adafruit/Adafruit_Learning_System_Guides/pull/204
slender iron
PSA: The SAMD51 datasheet was updated last month. Make sure you are using it.
idle owl
@tulip sleet Thank you
tulip sleet
@slender iron - they didn't change the dates on the datasheet page ๐ Errata is also updated. (I submitted a website bug report.)
datasheet page with obsolete dates: http://www.microchip.com/design-centers/32-bit/sam-32-bit-mcus/sam-d-mcus#mchp-Data Sheet
@slender iron @raven canopy In this new revB version of the samd51 datasheet:
tulip sleet
@slender iron do you have time for a quick consult before the mtg? could be in mtg video or discord. re event system
manic glacierBOT
slender iron
@tulip sleet sorry, was prepping for the meeting
raven canopy
@slender iron @tulip sleet thanks for the heads up on updated sheet. I may get a chance to peek at it tonight..
ruby lake
@errant grail I find the i2c OLED to be a bit slow when used with midi data. I'll have to try an SPI type. In my v1 module playground I use the toggle switch to turn off certain things like A/D and oled writes when faster midi actions are needed.
ruby lake
I'll have a look this afternoon
marble hornet
help needed with uploading to atsamd51. hey all, a few days ago i tried to install gdb, recommended by tannewt, so i can burn files to atsadm51 s. but i have not been able to figure out how. it seems like a good uploader. i am on a mac running 10.13.4 im fine with any program that will work to upload to atsamd21, 51, other common controllers. but i have been unable to get it to install. any and all help is appreciated
errant grail
@ruby lake I tested the TFT FeatherWing yesterday. It's much, much slower than the OLED in spite of its SPI interface. I believe that it's mostly due to the font bitmap to pixel routines that are coded in Python (and use the underlying vector graphics engine) and perhaps the framebuffer foundation. @slender iron suggested that improvements may be considered as the leeked game controller wing takes shape. Thanks for the idea of switching off the display during time-critical functions!
@stuck elbow Appreciate the I2C clock speed default suggestion for the OLED. I'll give it a try.
marble hornet
could you help @slender iron ? i used the JLink_MacOSX.pkg but it only installed the drivers. i think
slender iron
@marble hornet have you seen this guide I did? https://learn.adafruit.com/debugging-the-samd21-with-gdb?view=all#jlink-plus-metro-m0-express
its the same process for the 51, you just change the device id
marble hornet
@slender iron no i did not! thanks
slender iron
np ๐
stuck elbow
CircuitPython badge, latest prototype.
slender iron
<@&356864093652516868> anyone around to voice chat about error messages?
idle owl
Sure
manic glacierBOT
prime flower
@slender iron Sure, if channel is still going on this topic?
slender iron
ya
simple pulsar
hello
slender iron
idle owl
Hello @simple pulsar
simple pulsar
I have a quick question re: instructables.com - is there a preferred category for Adafruit / CircuitPython?
idle owl
I have no idea, I haven't dealt with instructables. Are you writing one up and wanted to know how to categorise it?
simple pulsar
Yes, i published one earlier. It has things like Arduino and Raspberry Pi. I ended up choosing Microcontrollers. I'd imagine not everything warrants a category but was just curious if this was something that had been discussed before?
idle owl
Not that I'm aware of, but that doesn't mean it hasn't been ๐
simple pulsar
Ok, if you spot anything you can add a reply to my recent post on https://forums.adafruit.com/viewtopic.php?f=60&t=136358 which mentions it, thanks.
idle owl
Ok!
slender iron
thank you for the help!
prime flower
thanks for letting me listen in
slender iron
np
idle owl
You're welcome @slender iron ๐
prime flower
@slender iron I just checked an AVR ASM book I have and it uses the same approach dan discussed
but it's weird, stores the sequence in LUTs
tulip sleet
idle owl
from adafruit_circuitplayground.express import cpx
cpx.play_tone(262, 1)
cpx.play_tone(294, 1)``` File "code.py", line 3, in <module>
File "adafruit_circuitplayground/express.py", line 617, in play_tone
File "adafruit_circuitplayground/express.py", line 643, in start_tone
File "adafruit_circuitplayground/express.py", line 598, in _generate_sample
TypeError: extra positional arguments given```adafruit_circuitplayground.express.__version__
slender iron
@tulip sleet looks like it works way better. thanks!
idle owl
@tulip sleet Did you add changes to the sound meter code in the CP Made Easy on CPX in the learn repo for 3.0?
tulip sleet
@idle owl I think so
I grepped for all calls that needed changes from frequency to sample_rate
idle owl
Ok.... I thought I forgot doing it, and was entirely confused...
idle owl
It didn't occur to me that any of my files would be in there because I haven't pub'd the guide yet. ๐
Fair enough ๐
I remember approving that now. I think I linked it in my mind with me updating the cpx lib so I didn't register that you'd updated any of my code.
And... now I have to rebase. ๐
slender iron
I changed mono to channel_count as well
idle owl
There's going to be another one to fix typos in the sound meter code but I have to get them up to the same point first.
Apparently. ๐
@tulip sleet Thank you!
simple pulsar
@idle owl hello - me again - qq on neopixels
Is https://learn.adafruit.com/circuitpython-essentials/circuitpython-internal-rgb-led complete wrt import statements for when the neopixel library is used? Would it not need the import line changed to import neopixel?
idle owl
Correct. That was caught a couple of days ago by someone who put in a PR for it but closed it for some reason. I have the updated code ready to go, but I didn't push it yet because I haven't had time to update the guide.
You do need to import neopixel for it to work.
simple pulsar
Thanks, that's what/ I just did on a CPE but was about to comment it out to see if it was imported by default super-magically/
idle owl
Nope. Good catch!
@tulip sleet https://github.com/adafruit/Adafruit_Learning_System_Guides/pull/207 and https://github.com/adafruit/Adafruit_Learning_System_Guides/pull/208
slender iron
keeps finding bugs
tough ridge
any good way to cut a list in half in python without doing the usual list[:length/2]? I'm realizing that method creates a duplicate which causes me to go over the memory limit (it's a big array)
idle owl
@simple pulsar Give the guide a refresh, and see if that works better for you.
@tulip sleet Thank you!
tulip sleet
@tough ridge do you need both halves? You can delete half the list with del l[0:length//2]
tulip sleet
it's pretty obscure!
slender iron
ugh, pulsein's interrupt handler called during tick handling:
#0 pulsein_interrupt_handler (channel=channel@entry=11 '\v')
at common-hal/pulseio/PulseIn.c:83
#1 0x0001c7d4 in external_interrupt_handler (channel=channel@entry=11 '\v')
at peripherals/external_interrupts.c:43
#2 0x0001d3b4 in EIC_Handler () at peripherals/samd21/external_interrupts.c:83
#3 <signal handler called>
#4 SysTick_Handler () at tick.c:45
@slender iron PDMIn bug manifestation seems to be that when things fail, the DMA transfers only one I2S word and then doesn't transfer any more. Trying to figure out what's wrong with DMAC at that point.
slender iron
weird
tulip sleet
yeah, I filled the buffers with sentinel values to see if they were being written at all and only the first word got overwritten
working on understanding the differences between dma startup in 2.x and 3.0
idle owl
@tulip sleet Sorry... there's another one https://github.com/adafruit/Adafruit_Learning_System_Guides/pull/209
idle owl
@tulip sleet Thank you
tulip sleet
@slender iron I am working on the rotary algorithm, cause I think the version as coded is still potentially bouncy. Will give you a sample impl in Python tonight or tomw mornign.
slender iron
yeah, I realize it might skip states and be incorrect but didn't think it was a huge issue
how about doing the review besides that and following up?
surreal saffron
Thoughts on changing all the "cannot"s in error messages to "can't"s? I was looking for a way to free up enough space to do strerror for errno on gemmas, and it seems like a simple and obvious way to free up some room... At the risk of merge conflicts for pulling stuff from MicroPython in the future.
Or, in general, thoughts on freeing space by rewording messages to be more succinct?
I have the cannot -> can't one as a PR I could send... I just waffled on it because I wasn't sure how you'd feel about the unnecessary divergence from MicroPython.
slender iron
@surreal saffron How much space does it free? Ideally we'd find something with a larger payoff.
stuck elbow
.oO( we have lzw compression implemented... )
slender iron
implemented where?
stuck elbow
for upip
at the least
and I'm thinking displaying an error message doesn't have to be fast
stuck elbow
nice
tulip sleet
@slender iron ok, made a rotaryio PR on your own PR, via your personal repo
tulip sleet
@slender iron if you want to wait you could finish your PR and I'll submit a PR to adafruit/circuitpython with the encoder changes
i don't understand why pirkey build is failing, something cached??
reef seal
requesting a code review: https://github.com/adafruit/CircuitPython_Community_Bundle/pull/9 . @slender iron @idle owl
surreal saffron
The cannot -> can't change frees up only 16 bytes on Gemma. That happens to be enough for a change I want to make to get errno names on Gemma - still no description, but at least "EEXIST"
errant grail
I've looked around but can't find an answer to this question: When creating a list for the MIDI protocol stack controller definitions (there are 127 of 'em), what's the most memory efficient way to create blank entries in the list? MIDI has quite a few undefined controllers that are inserted into the stack definition in non-contiguous blocks. I tried using a dictionary but it wasn't as space-efficient as the 127-element list. Here's an example taken from the middle of the list:
('NRPN_MSB', 'Value'),
('RPN_LSB', 'Value'),
('RPN_MSB', 'Value'),
('', ''),
('', ''),
('', ''),
('', ''),
('', ''),
('', ''),
('', ''),
('', ''),
('', ''),
('', ''),
('', ''),
('', ''),
('', ''),
('', ''),
('', ''),
('', ''),
('', ''),
('', ''),
('All_Sound_Off', 'Value'),
('Reset_All_Ctrls', 'Value'),
('Local_Ctrl', 'Value'),
('All_Notes_Off', 'Value'),
('Omni_Off', 'Value'),
('Omni_On', 'Value'),
('Mono_On', 'Num_of_Chans'),
('Poly_On', 'Value'),```... I'm using the list offset to match the MIDI controller code (0 - 127).
tulip sleet
@errant grail Do you need these strings to show on a display or something else human-readable?
errant grail
Yes. It's working now for my MIDI sniffer, but I'd like to add more features and am running out of memory. It's particularly problematic given the memory needed by different display libraries.
I've shortened the descriptors in the table and that seems to help, so I'm guessing that I'm on the right path.
tulip sleet
you could use Camel case to avoid the underscores
multiple instances of identical strings are no duplicated, so for instance 'Value' is only stored once
errant grail
Good idea about the Camel case. I tried to stay close to the MIDI standard defs where possible, but all bets are off now. Good to know that identical strings aren't duplicated. Whew.
tulip sleet
you could store that table as a byteaarray where the values of the bytes point into a list of string constants. you could also not use a nested list, but instead stored two parallel lists or just a single list where the even items are the first elemetns and the odd items are the second
errant grail
I also found that an unused entry in the table doesn't need a tuple.
tulip sleet
but lots of strings just takes space, period
errant grail
Yeah. I think the next step would be to split the table to avoid the non-defined gaps. It would be pretty easy to provide an offset translation in the code.
All great suggestions, Dan! Thank you! I have a few more forks in the rabbit hole to investigate now.
tulip sleet
you could do string expansion at runtime for common substrings, like wherever you have 'Ctrl', use '^' or some other character, and then substitute only when you need the string temporarily: 'Reset_All_^s' and 'Local_^'. ... The gaps in the table don't use up that much space, it's the strings themselves I think that are the bulk.
errant grail
Oh, that would help, indeed. There's a lot of repetition in the control codes table.
errant grail
I like the file idea, but am also wrestling with response time to try to keep things real-time with the MIDI events. I'll try it and compare. It may be the better trade-off if I switch to sampling instead of real time sniffing. BTW, the rest of the MIDI protocol stack is pretty simple.
slender iron
@errant grail how much empty space is in the array? A dictionary might actually be better
ints can be keys in a dictionary
@tulip sleet back now, was doing errands
errant grail
The list is about 20% empty. I tried the dictionary approach using a byte as the index, but it was about as inefficient as the 128-element list approach.
surreal saffron
Does the CircuitPython atmel-samd port use lwip or axtls?
surreal saffron
Great
errant grail
@tulip sleet @slender iron You've given me some good ideas so rather than dragging you further down this hole, I'll go back and experiment. I like the string expansion and file approaches. Oh, and I should compare beta 3.0's memory management scheme to see if that helps. Thanks!
slender iron
@errant grail you could also skip the tuples and index [2*x] instead
errant grail
Can you share an example of that? I'm still a neophyte...
like in two separate lists?
are tuples memory hogs?
surreal saffron
How about libffi? atmel-samd use that?
slender iron
@errant grail your example would be:
['NRPN_LSB', 'Value',
'NRPN_MSB', 'Value',
'RPN_LSB', 'Value',
'RPN_MSB', 'Value',
``` etc@slender iron ah, got it.
slender iron
then the first value is foo[2*code] and the second is foo[2*code + 1]
@surreal saffron I don't think so
errant grail
very comma-delimited-like
tulip sleet
@surreal saffron just checked no
why are you asking? we have no network support in the atmel-samd ports
only in esp8266 port
surreal saffron
Double checking my assumptions.
errant grail
@slender iron also lends itself to Dan's suggestion to keep the verbose descriptions in a file.
tulip sleet
nrf port (in progress) uses libffi
surreal saffron
The uerrno module lets ports define what errnos they can raise. I want to prune that list for atmel-samd
errant grail
Okay off to the rabbit hole. Thanks again!
tulip sleet
@surreal saffron makes sense - thanks for looking at that!
surreal saffron
I've had it half finished since PyCon, heh...
tulip sleet
@slender iron have to be out for half an hour or so back soon
slender iron
kk, I'm code size hunting
tulip sleet
still debugging PDMIn. Something similar to this line is in the 2.x impl:
// Force I2S to start.
I2S->DATA[0].reg = I2S->DATA[0].reg;
adding that to the 3.0 port makes the PDMIn loop test run for much longer, BUT then there's MSC activity, and I get a hard fault when it's doing an spi read (sync, not dma). looks like the spi device descriptors were smashed
maybe it's all buffer overruns somewhere
slender iron
ยฏ_(ใ)_/ยฏ
if you need a break you could look at another issue
the pdm one should be fixed by 3.0 but we have time
rigid path
Might be a dumb question, but can I use 2 feathers (not wings) together on a feather tripler board? I am playing with a pair of Ferather M0 RFM69HCW Rasdio Boards.
On the transmit end, I have a feather tripler board loaded with an M0 Packet Radio, an Ultimate GPS Feather Wing and an OLED Feather Wing.
On the receive end, , I have a tripler board holding an M0 Packet Radio and an OLED Feather Wing. Right now, everything works fine, the GPS info is being transmitted to the receiver and displayed on both OLEDs just fine.
On the receive end, I want to replace the OLED feather wing with an M0 Adalogger to record the GPS readings. My question is... can an M0 Adalogger and an M0 Packet Radio Feathers co-exist on the same tripler and work together?
slender iron
@rigid path I don't think so since they'll both try to power it.
rigid path
@slender iron - I figured that might happen. Looks like I need to pick up an Adalogger feather wing. Not such a bad thing. ๐
slender iron
๐
surreal saffron
Is lib/berkeley-db-1.xx linked in on atmel-samd?
If not, I think we'll never raise EPERM...
manic glacierBOT
tulip sleet
@surreal saffron it's used on the esp8266 port. It's used if MICROPY_PY_BTREE is 1
stuck elbow
why do we even have the database stuff enabled
surreal saffron
Ok - likely to ever be used on atmel? If so I'll leave it in the list of errnos we can raise, if not I'll remove it.
stuck elbow
can you make it conditional on the same variable?
surreal saffron
no, not really...
faint summit
I'm running into some basic issues with mpy files on my metro m0. Is it alright if I ask them here?
idle owl
Yep, ask away! Someone will answer when they can
surreal saffron
@stuck elbow actually scratch that, i can - though it's a bit ugly to do so...
faint summit
Thanks! I want to get a .mpy of my own module so I pulled from the micropython repo and ran mpy-cross. I moved the mpy file into my lib folder and when I try to compile, I get the error "ValueError: Incompatible .mpy file. Please update all .mpy files. See http://adafru.it/mpy-update for more info."
I read through the site and I think it's referring to their own libraries? Not too sure what's going on tbh
idle owl
That happens when you run a version of mpy-cross built with a different version of CircuitPython than you're running on the board.
So if you built mpy-cross from a 3.0b CircuitPython, and you're trying to use the .mpy file on 2.x CircuitPython, you'll get that error. Or vice versa.
You said you pulled mpy-cross from the micropython repo? Do you mean CircuitPython?
faint summit
hmm I pulled from here https://github.com/micropython/micropython
surreal saffron
will extmod/machine_i2c.c be linked into the atmel-samd port? (How can I figure this out myself? :D)
idle owl
@faint summit Ok that might be the problem. You need to use mpy-cross from CircuitPython with CircuitPython
slender iron
@tulip sleet my local build of pirkey is out of space as well
tulip sleet
@slender iron I'm not seeing that, but I will reclone your repo and try again
tulip sleet
yah i did. new clone fails. will diff the two trees
i didn't change anything that should affect that build in my PR to you (famous last words)
faint summit
@idle owl d'oh I see. I'm running on the 3.0.0 beta but it looks like the mpy-cross is behind? (https://github.com/adafruit/circuitpython/tree/master/mpy-cross) Should I just go back to 2.x?
idle owl
@faint summit You can build it yourself if you clone the repo locally. It's a few steps, it's not bad. What OS are you running?
faint summit
macos high sierra
tulip sleet
@surreal saffron you can just look in the ports/atmel-samd Makefile and/or grep for that file in the atmel-samd tree.
surreal saffron
cool - wasn't sure if just the Makefile was enough or there were other places to check...
tulip sleet
@slender iron ok yes my submodules were out of date but not in an obvious way. Two of the submodule dirs were empty
is rotaryio turned on for non-express?
more likely it's the slight enlargement of the eic stuff
slender iron
it shouldn't be and I couldn't find it in the map
the busio constructors take a bunch of space
tulip sleet
yep, that was my motivation for removing them when we had that discussion. ifdef'ing them out is a pain
or reducing them to NotImplemented by ifdefs
slender iron
right
another idea I had was to shrink the mp_obj_type size
there are ~120 instances of it in memory
at 60 bytes each
tulip sleet
setting inline-limit to 50 instead of 60 makes pirkey fit with 380 bytes free
i don't see anything you can remove from mp_obj_type. you could put the optional stuff at the end ordered by priority and have a set of flags saying whether they exist or not, shortening it in some cases. or a table of 4-bit offsets saying where to find each of those function pointers in a block of function pointers, if they exist. Sounds like a 4.0 project. everything else is fine except pirkey so we can just squeeze that however we can
we're out to eat now for a couple of hours or so, will check back when I return
slender iron
k
faint summit
@idle owl Alright, thanks so much! Sorry for making you spoon feed me
idle owl
@faint summit Don't be sorry! We all started somewhere. When you're ready, I can show you how to build it yourself ๐
prime flower
is TouchIO incl'd in the crickit CPX build?
idle owl
It's always clocks.
slender iron
today is clock day too ๐
cunning crypt
Just don't get ticked at them.
idle owl
The first iteration of the new Adafruit CircuitPython Code of Conduct. Once completed, it will replace the current Code of Conduct displayed on Discord, and apply to all CircuitPython spaces. Please read through it and provide any suggestions you may have as a review or as a comment on the pull request. We want this to be a great fit for our community, and the best way to do that is to have community input. Everyone is welcome to submit a review for consideration. https://github.com/adafruit/Adafruit_CircuitPython_Code_of_Conduct/pull/3
cunning crypt
Being a CP Code of Conduct, would it replace #code-of-conduct or no?
Just so we're perfectly clear
cunning crypt
Just starting to read everything now
grumbles about being a writing-minded person
There's a few places that are perfectly fine, but there's better wording.
idle owl
So submit a review for consideration ๐
The wording is a bit awkward in general because it's a Code of Conduct.
cunning crypt
Oh, yeah
cunning crypt
But I have to finish reading the whole thing before making any sort of review
idle owl
That's fair
idle owl
@slender iron My notes are evidently sparse on getting a new repo ready for the bundle. I think I took better notes on RTD than I did Travis. I don't remember how to run Adabot (all of that was on my previous machine, so I'm assuming it didn't occur to me to take better notes on it). I activated it on Travis (I think) but "trigger build" says "custom build will not appear on GitHub" so I'm not sure that's correct. Do you have a few minutes?
slender iron
@idle owl instructions for adding something to the bundle is here: https://github.com/adafruit/Adafruit_CircuitPython_Bundle#adding-a-library
do need to know more?
idle owl
I'm still at the bit where my notes say run Adabot, find Travis, activate, trigger guild. Then RTD is next following that. The actual adding to the bundle part, I knew was documented.
cunning crypt
@idle owl Regarding your CoC - I feel like it should apply to all Adafruit github repos, but that seems like a little bit of a stretch for a CircuitPython CoC
idle owl
We already discussed that. The Adafruit Be Excellent policy already covers Adafruit spaces. This is being written for CircuitPython spaces.
cunning crypt
I just like unified systems, so like to cover everything with as few things as possible
idle owl
Understandable. But in the end, not intended to be our scope. Adafruit has Adafruit covered. We're working to outline what we're doing here. We already had two separate Codes with the current Discord #code-of-conduct and the Adafruit Policy. So this isn't a change to that.
cunning crypt
Entirely fair.
cunning crypt
@idle owl There, did the review. Mostly pedantic stuff, if I'm being entirely honest
faint summit
@idle owl Hopefully I get there quick enough ๐
@idle owl I actually have another question if you don't mind
cunning crypt
@faint summit Just go ahead and ask the question. Kattni can answer it when she can, or someone else might have an answer quicker.
faint summit
Alright! I'm wondering how I should go about with large lists (like thousands of entries). I tried converting my own module into a .mpy format, but I'm still getting memory allocation errors.
cunning crypt
List as in the variable type, or list as in a text file that's loaded? Either way, they take up a lot of space compared to the device's RAM
A huge list might not be possible to fit on the board.
faint summit
as a variable type. I'm using lists of ~100 entries, and I definitely need much larger ones
raven canopy
@cunning crypt sweet! a partner in pedanticizing on the CoC! ๐
cunning crypt
@faint summit Metro M0 Express, right?
faint summit
@cunning crypt yep! Ok, maybe not that large, but is it feasible for me to get something like 1000>
cunning crypt
The M0 has 32k of RAM.
That's... 32768 bytes
Not counting what Circuit Python uses by itself
surreal saffron
or the overhead of each python object
cunning crypt
You can fit around 3275 total variables with a character length of 10 each
Disregarding all other uses of RAM - Most significant of which will be Circuit Python, but as @surreal saffron mentions, the program itself will eat away at it too
faint summit
Ok, then what are some steps I can do to maximise the list size I can use?
slender iron
what are you storing in it?
cunning crypt
Make the length of each individual object as short as possible
You can store twice as many items with 5-character names as you can with 10-character.
surreal saffron
use fewer objects, if possible? You can fit more in a single big string than in 10x as many strings that are 1/10th the length
cunning crypt
Each string has an end character, correct?
So that's an extra character per string
faint summit
just a lot of floats
cunning crypt
(insert "Whatever floats your boat" joke here)
faint summit
the goal was to get lots of readings so I can take a reliable fft
cunning crypt
You probably don't need that many.
surreal saffron
@cunning crypt I don't believe they have a terminator, but they do each have a reference count, and a length, that are stored as part of the object
@faint summit Don't use a list, use an array - http://circuitpython.readthedocs.io/en/2.x/docs/library/array.html
without any doubt, the maximum number of floats you can fit in memory will be a single array of floats.
faint summit
@cunning crypt fair enough, I did some quick testing and realised that I need >400 readings.
cunning crypt
@faint summit If you can't fit everything into the M0
The Metro M4 has 6 times as much RAM
idle owl
@cunning crypt Can you submit an actual review? You can click on line numbers and do in-line change-requests - it allows for discussion on each one, and also keeps track better as changes are made. If you're unsure how to do that, I can help you with that. (Not sure if you chose to do a comment or if it wouldn't let you directly review it.)
cunning crypt
@idle owl I did what was the most obvious and seemed right. I'm not familiar with GIT reviews at all.
idle owl
@cunning crypt Would you like to learn?
cunning crypt
That's a silly question. Of course!
idle owl
๐
faint summit
@cunning crypt that looks promising. Do you think this is a realistic fix to this problem?
cunning crypt
@faint summit I'm fairly certain the issue you're running into is memory. But as @surreal saffron mentioned, try using an array first to see how that goes.
surreal saffron
try this in a repl:
import array
a = array.array('f', range(4000))
that works for me, and creates an array of 4000 floats
you can access them with a[0], a[1], etc
for both read and for write.
idle owl
Ok, first go to the pull request. Then click on "Files changed". Then scroll down to the code-of-conduct.md file in that list. You'll see if you mouse over any given line in that file, a blue [+] symbol will appear on the left side between the line number and the text. If you click that plus symbol, it opens a little review window. For the first one, you type in your line specific comment there, and click "Start a review". It will save it and say "Pending". Then go to the next line you want to do that, and repeat, click the +, enter your comment. This time, click "Add review comment" to continue. Once you've finished adding all your in-line suggestions, you go up to the top and click "Review changes" on the top right. Enter some quick text there, and then choose "Request changes" from the list under the text field there, and then click "Submit review".
cunning crypt
Ooh, got it!
idle owl
You'll see some inline comments there already from Scott. You can still add yours if it's the same line if he commented on something that you also want to comment on. And that's what it will look like once you're finished submitting the review.
cunning crypt
@idle owl Done
idle owl
Beautiful, thank you!
surreal saffron
can anyone more familiar with CircuitPython's tests tell me what the heck https://travis-ci.org/adafruit/circuitpython/jobs/386912838 is trying to tell me?
it looks like it failed to find the expected test output - so that's weird.
raven canopy
@surreal saffron expand the "Building Adafruit binaries" section. quite a few errors.. looks like mostly "first use"
surreal saffron
aha.
that's probably just a header that needed to be included and wasn't, i'm guessing.
thanks
raven canopy
yeah, with that many, i would start there too. you're welcome!
idle owl
@slender iron I responded earlier, but I don't know if you missed it. The documentation you linked me to isn't the issue I'm having. I'm still trying to get travis setup on the new repo. The repo exists already and has what appears to be all the right files in it. I activated it in travis (I think, I found the repo and clicked activate). I have "Run Adabot" in my notes before the travis stuff though. Is that to try to get Adabot to do the travis stuff for me? I don't remember how to run Adabot. The thing in my notes after "Activate Travis" is "Trigger build" - I found something that says trigger build, but it says it's custom and won't be linked in GitHub so I wasn't sure that was correct.
idle owl
Oh under Environment Variables?
slender iron
yeah, so it looks like it hasn't added it
idle owl
I didn't run it. So I guess that would make sense.
slender iron
running adabot is python -m adabot.circuitpython_libraries
it runs every night too but depends on permissions being correct
er, not every night. every time I do it for the meeting
idle owl
I clone the repo and run that inside the repo?
slender iron
oh thats right, new comp
idle owl
Yeah... :/
slender iron
yeah, thats the start. then you need to copy and edit the env.sh template to include your permissions keys
idle owl
Did you give me permission keys last time? Because it doesn't look like I have any generated.
@slender iron What kind of scope am I supposed to give the key?
idle owl
that sounds accurate
I feel like I remember that part of the discussion.
Because you said later sometimes it asks for them
slender iron
right
idle owl
Do you want to do that again or no
slender iron
I don't know. I'm busy poking rosie and am about to peace out for the weekend. can we do it monday?
idle owl
Yeah, sorry about it trying to get to it so late.
slender iron
np
tulip sleet
@slender iron if you aren't trying to get the latest PR to build I'll wait. CPX build is slightly too big. I can make LIS3DH library smaller by at least 300 bytes.
slender iron
I hadn't seen it. I'll look monday
tulip sleet
kk - I'll look at reducing library sizes in a more orderly way.
slender iron
ideally we'd leave some extra space so we're not always fighting it
raven canopy
git continues to amaze me. i'm rebasing WIP with the updated SAMD peripherals file/function setup, and its going pretty seamlessly. ๐ ๐
fading solstice
Anyone working on or know of someone that is working on: a CircuitPython driver for L3GD20 Triple Axis Gyroscope.
idle owl
@fading solstice Not that I'm aware of. If you're asking because you'd like to, file an issue on the CP repo so others know and so we can track it. Also check the issues list for it if you haven't already.
idle owl
I don't think anyone is working on it. So that would be great ๐ Please update the issue and assign it to you if you decide to work on it.
fading solstice
thx
idle owl
Thanks for looking into it!
stoic gazelle
I'm looking for the order of miso, mosi, and sck on the metro m4 express, I see them on the pinout page but the page doesn't specify the order of pins
tulip sleet
@idle owl would you be willing to approve https://github.com/adafruit/circuitpython/pull/895?
@stoic gazelle it's the same as on the Metro M0 and the Arduino. I'll look for a picture for you.
idle owl
@tulip sleet I was going to ask if it was something I could do. I'll look at it.
tulip sleet
@stoic gazelle note that the VCC pin is 5V, not 3.3V, even on the M0 and M4. But the SCK, MOSI, and MISO pins are 3.3V logic: don't apply 5V to them.
@idle owl Thanks!
stoic gazelle
Ah ok thanks, I am connecting the m4 to a project I started with a feather on a tripler(and a doubler)
stoic gazelle
The hope is to use the joy featherwing with the GPS feather with and the SD writer to ba able to track certain things - my purpose is to track potholes so I can send the city/county a list rather than every so often give them a call with only a rough idea of where the potholes are
So far I am just past where I was with the m0 feather
stoic gazelle
Tried putting the GPS antenna by the window but it is cloudy out so I'll try again later
tulip sleet
@stoic gazelle clouds don't affect GPS signals in any significant way
stoic gazelle
lol then I should try outside
Thanks!
I put the end of the antenna out the window but I still don't get a fix, are there other things I can do to get it working?
tulip sleet
I've gotten a fix just by hanging it in the window, so there may be some other issue. If you just remove the antenna and keep it outside, do you get a fix? Are you getting reasonable responses back from the GPS to the microcontroller board?
you should not need an external antenna outside if you have a clear view of the sky
the GPS module has an antenna built into its cover
stoic gazelle
I used the example code and it is responding with: waiting for fix
Ok, when I head out I'll bring a laptop so I can check it outside
tulip sleet
you may have tx and rx reversed going to the module. I believe you'll see that if that's true.
rx->tx and tx->rx is what it should be
stoic gazelle
Oh I'll try that first
tulip sleet
if you're using a feather doubler/tripler then the wiring is automatically correct from feather to GPS featherwing
stoic gazelle
I switched the wires and get the same results, ah okay I'll switch back :)
tulip sleet
From the guide https://learn.adafruit.com/adafruit-ultimate-gps-featherwing/overview: The LED blinks at about 1Hz while it's searching for satellites and blinks once every 15 seconds when a fix is found to conserve power.
that's the LED on the GPS featherwing
stoic gazelle
After switching the wires back to how they were I now have GPS output!
tulip sleet
maybe they were just loose or something
tulip sleet
congrats!
stoic gazelle
Thanks to @idle owl for suggesting to switch to the m4!
idle owl
@tulip sleet If you have a minute, could you approve this for me please? https://github.com/adafruit/Adafruit_Learning_System_Guides/pull/211
raven canopy
so rebasing wasn't as seamless as i thought... still better than re-writing from scratch, though. ๐
manic glacierBOT
Well, good to know that correct bootloader and cables can be ruled out. I think. ๐
I just loaded Arduino IDE. Had to update my BSP to 1.2.0. I did get the "Warning: non trusted" message after the update ran. Closed the IDE, re-opened (normal, non-admin), and the M4 boards showed up for me. Also on Win10, with the Windows Store version of the IDE.
Didn't attempt an uploa...
stuck elbow
Raspberry Pi is a computer, not a microcontroller
MicroPython is Python for microcontrollers, and it has been ported to many architectures: STM32, PIC32, ESP8266 and ESP32, NRF5x, SAMD and many more.
there is an incomplete list of boards on the wiki: https://github.com/micropython/micropython/wiki/Boards-Summary
stuck elbow
no
timber mango
More coder friendly?
@errant grail on small platforms I think of strings as 'media' (just like JPEGs and WAVs and other audio/video files that contain digitized imagery and sounds).
wrt your discussion about the string table with the gaps for your MIDI stuff.
stuck elbow
@timber mango easier, yes
stuck elbow
really depends on your program
SAMD21 has 32kB and that's a bit tight for larger programs
errant grail
@timber mango Yes -- it's a measured trade-off between using default data handlers (arrays, lists, dictionaries) and indexing streams. Because I prefer to have verbose data attributes, it's fun to find ways to compress and expand the information, not unlike dealing with media streams through limited bandwidth. Dan and Scott gave me some excellent ideas to pursue.
marble hornet
DS3231 HELP. hey, i'm having trouble import the ds3231 library. i downloaded the file and copied it into the lib folder. i tried in the repl: from adafruit_ds3231 import DS3231 but i got this error:
File "<stdin>", line 1, in <module>
ImportError: cannot import name DS3231```
this isn't ive seen the 'can't import name error.' my googleing of that error doesnt yeild results like when i google other errors.
does anyone know what is causing this????thank you
raven canopy
@marble hornet try just import adafruit_ds3231. DS3231 is the only defined class, with no additional inheritances which may be causing the error.
raven canopy
yw! i wish i could give a better explanation as to why/why not. i would just be regurgitating information from the pydocs...
marble hornet
wait
@raven canopy dir(adafruit_ds3231.DS3231) says that adafruit_ds3231 doesn't have attribute DS3231
to me, it seems like the same problem
raven canopy
what does dir(adafruit_ds3231) output?
marble hornet
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
AttributeError: 'module' object has no attribute 'DS3231'```hmm...
marble hornet
that is after: import adafruit_ds3231
raven canopy
i'm looking right at in the repo... ๐
give me a sec, and let me get a board running the latest libraries.
marble hornet
ill try re-downloading
raven canopy
yeah, something may have gone wonky. here is what i get:
>>> import adafruit_ds3231
>>> dir(adafruit_ds3231)
['__repo__', 'i2c_bcd_datetime', '__name__', 'i2c_bcd_alarm', 'DS3231', 'i2c_bit', 'I2CDevice', '__version__']
>>> dir(adafruit_ds3231.DS3231)
['__module__', 'alarm1_status', '__qualname__', 'alarm1_interrupt', 'alarm2', 'alarm2_interrupt', 'datetime', '__init__', 'alarm1', 'disable_oscillator', 'datetime_register', 'alarm2_status', 'lost_power']
and this:
>>> from adafruit_ds3231 import DS3231
>>> dir(DS3231)
['__module__', 'alarm1_status', '__qualname__', 'alarm1_interrupt', 'alarm2', 'alarm2_interrupt', 'datetime', '__init__', 'alarm1', 'disable_oscillator', 'datetime_register', 'alarm2_status', 'lost_power']
so the single class thing i said earlier: doesn't make a difference after all. i understand less python than i let on. ๐
marble hornet
@raven canopy its working now!!! must have been a bad download
thanks for the help
raven canopy
yay! and not sure i helped all that much. hehe
marble hornet
@raven canopy it's the thought that counted, do you know if extended ram is supported yet in alpha ? or beta ?
raven canopy
extended RAM? as in, the new heap management (yes)? or using an external source for RAM (don't know that will ever happen)?
stuck elbow
esp32 can use external ram, but samd doesn't have such possibilities
marble hornet
hi! how does one get info out of a time_struct ? thanks for help either way!
raven canopy
@marble hornet time_struct is a named tuple. this is from the ds3231_simpletest.py on github:
while True:
t = rtc.datetime
#print(t) # uncomment for debugging
print("The date is {} {}/{}/{}".format(days[int(t.tm_wday)], t.tm_mday, t.tm_mon, t.tm_year))
print("The time is {}:{:02}:{:02}".format(t.tm_hour, t.tm_min, t.tm_sec))
time.sleep(1) # wait a second
@raven canopy thank you, and sorry. i didn't think to check there. my bad.
raven canopy
here is the parameter list: https://circuitpython.readthedocs.io/en/latest/shared-bindings/time/__init__.html#time.struct_time
and no worries! i'm used to looking everything up, so its become second nature. ๐
marble hornet
WHOO!! i just got threading of tasks working on circuit python based on assigned priority !!!!!!!!! yay
also, when you have been coding for so long in cir-py you do ctrl s instead of enter to send literally this
raven canopy
heh. yeah, i type Markdown all over the place now, thanks to GitHub and Discord... _ instead of ctrl+i.
marble hornet
markdown?
@raven canopy
also #running_out_of_memory oh yay!
another question:
what would happen if i put main.py in the frozen modules ?
idle owl
The planet would implode.
cunning crypt
We wouldn't recommend it. Un-imploding the planet is a bit difficult, as you may imagine.
stuck elbow
too late: https://qntm.org/board
cunning crypt
I learned a new word: Geocide
idle owl
So you did it?
What happens if there's one in the frozen modules and also one locally, which one does it prefer?
That's entirely curiosity, not that I'm implying an issue.
cunning crypt
So, having a frozen main.py could potentially create conflicts.
Having a non-frozen one lets you ensure that the user doesn't accidentally cause issues.
tulip sleet
@marble hornet instead of freezing main.py itself, you could just have it import a module and then call a function in that module (or just have the import actually run the program). In other words you have added your own privately library.Then you don't usurp main.py. Or if you are using existing Adafruit libraries that are large, you can freeze those to save RAM. They are much less likely to change than your own program, so freeze them first.
marble hornet
@tulip sleet wait what! frozen modules save ram! is that b/c they are .mpys ?
tulip sleet
.mpy's from the file system still need to be copied into RAM to be used. But frozen .mpy's are already in the address space (in flash) so they don't need to be copied. The objects they create of course live in RAM, but the code itself stays in flash, which can be a real RAM saver.
marble hornet
what! thank you
astral vessel
Hmm, maybe I'm being blind here. I'm connected to a trinket M0 using screen and the print() command running in main.py works fine. But I want to send data the other way, I want to type into my screen session and for those characters to be saved to a variable by the main.py script. So something like uart.read() but for data coming over the usb connection. (To be clear I am not, and do not want to be, using the REPL)
I'm assuming it's something simple I'm missing
astral vessel
Why would I type my string into the function? It should be polling the usb connection and getting whatever characters I type there.
ah, I see, that's gets printed
marble hornet
sorry @astral vessel i should have used a better string
astral vessel
Now if I'm in the main while loop, input() is going to block until it receives a newline. That's not very useful behavior. Is there a way to check for data in a buffer or something and act on it when there is? Like the behavior in the UART example here: https://learn.adafruit.com/circuitpython-essentials/circuitpython-uart-serial
tulip sleet
@astral vessel That is a deficiency we've discussed but not solved yet: https://github.com/adafruit/circuitpython/issues/231
astral vessel
ah, well fortunately I can work around that
idle owl
@tulip sleet My sound_level code is still hanging the filesystem. But it all imports ok and play_sound works separate from sound_level. I'm freezing it into a custom build to see if it works from there or not.
That's a new one. ```-332 bytes free in flash out of 253440 bytes ( 247.5 kb ).
4084 bytes free in ram for stack out of 32768 bytes ( 32.0 kb ).
Too little flash!!!```
No idea what to do with that.
astral vessel
New question: I have an esp connected to the uart and am reading in the data from the esp when it lists the available networks and printing that data back to my computer, but it's too slow and misses data. Suggestions?
while True:
data = uart.read(32)
if data is not None:
data_string = ''.join([chr(b) for b in data])
print(data_string,end="")
I have a weird problem
a really weird problem
im running this:
class thread_list():
def __init__(self, length = 3):
self.thread_list = [[]]*(length+1)
def add_task(self,func, arg_tup, priority):
self.thread_list[priority].append(task(func, arg_tup))
def chug(self):
for cur_list in self.thread_list:
while len(cur_list):
cur_list[0].perform()
cur_list.pop(0)
def __init__(self,func, arg_tup):
self.func = func
# set arg_tup as a tup whether or not inputed as tup or not
if type(arg_tup) == tuple:
self.arg_tup = arg_tup
else:
self.arg_tup = (arg_tup,)
def perform(self):
self.func(*(self.arg_tup))```and when i make a thread_list object and add_task to it the newly created task is added to all the lists inside of the thread_list object. anyone know why, am i missing something ?
ahhhhhhhhhh
ANY help is greatly appreciated
here is some tested code:
>>> from tg_modules.tasking import thread_list
>>> foo = thread_list()
>>> foo.add_task(print,('this peice of text was added first'),3)
>>> foo.add_task(print,('this one was added second'),1)
>>> print(foo.thread_list)
[[<task object at 20001440>, <task object at 20001470>], [<task object at 20001440>, <task object at 20001470>], [<task object at 20001440>, <task object at 20001470>], [<task object at 20001440>, <task object at 20001470>]]
>>> ```why does any given task (say the one here: 20001440) appear more than once.? i'm assuming i made a mistake.
THANKS FOR LISTENING ! ANY AND ALL HELP IS MUCH APPRECIATED
@raven canopy ?
i will be back tomorrow too
raven canopy
@marble hornet the problem looks to be the empty list(s) at the start. I ran this in python3:
>>> foo = [[]]*4
>>> foo[0].append([0])
>>> foo
[[[0]], [[0]], [[0]], [[0]]]
>>> del foo
# set an item in the list before appending
>>> foo = [[]]*4
>>> foo[0] = [0,]
>>> foo
[[0], [], [], []]
>>> foo[0].append([1,2])
>>> foo
[[0, [1, 2]], [], [], []]
>>> foo[1].append((0,1,2))
>>> foo
[[0, [1, 2]], [(0, 1, 2)], [(0, 1, 2)], [(0, 1, 2)]]
I didn't see anything in the pydocs that would explain it, but i didn't look beyond the mutable sequences page...
it seems like a slightly kludgey workaround, but i changed add_task to this:
def add_task(self,func, arg_tup, priority):
if self.thread_list[priority]: # edit: dropped use of len(), which is less pythonic
self.thread_list[priority].append(task(func, arg_tup))
else:
self.thread_list[priority] = [task(func, arg_tup)]
and it worked, like this:
>>> import tg_thread_test
>>> foo = tg_thread_test.thread_list()
>>> foo.thread_list
[[], [], [], []]
>>> foo.add_task(print, ('test1'), 0)
>>> foo.thread_list
[[<tg_thread_test.task object at 0x000002166C24A208>], [], [], []]
>>> foo.add_task(print, ('test2'), 1)
>>> foo.thread_list
[[<tg_thread_test.task object at 0x000002166C24A208>], [<tg_thread_test.task object at 0x000002166C24A240>], [], []]
>>> foo.add_task(print, ('test3'), 0)
>>> foo.thread_list
[[<tg_thread_test.task object at 0x000002166C24A208>, <tg_thread_test.task object at 0x000002166C1D0A90>], [<tg_thread_test.task object at 0x000002166C24A240>], [], []]
>>>
@slender iron I won't be able to make it to the meeting tomorrow. My "out Thursday" turned into "out Thursday+Friday, and Mon-Wed" with 30+ hours of combined driving and hotel life for 5 days. My brain is mushy. ๐ต Here are my hugs-n-status.
Hugs: @slender iron for making the SAMD peripheral functions more modular; really opens up some future growth. @tulip sleet for the myriad of troubleshooting on GitHub & Discord (and probably the forums, too) this week. @errant grail for the MIDI boards and exploration; its inspirational to see that stuff pop up. And a group hug for all the things my exhausted mind can't conjure up right now.
Status: didn't get much further on FrequencyIn. Did manage to rebase to current master with the peripheral function changes. That was my mushy brain limit. Will re-attack when I get back later this week.
totally didn't want that to ping everyone... ugh.
slender iron
@raven canopy thanks! added to my notes
<@&356864093652516868> and everyone else. Our meeting will happen as normal tomorrow here on Discord. It'll be 11am Pacific / 2pm Eastern.
manic glacierBOT
marble hornet
@raven canopy your altered add_task method worked perfectly. thank you. i'm not 100% sure that that difference makes it better but i see how it could be used again. i would like to give you a line of credit in the code, what should i put for your name?
umbral dagger
@marble hornet When you make the list and give it an initial element value, youโre saying โCreate a list that long and put this object in each cellโ. In your case โthis objectโ was an empty list, so each cell got the same empty list object. So when you appended to one, since they were all the same list, it showed up in each place. @raven canopy โs code will create a new sublist each time one is emptied and a new task is added. Might not be a problem but worth being aware of.
@slender iron Back from vacation (with a cold ๐ฆ ) and with a growing focus on CP. Will aim to make the meeting today!
solar whale
Does anyone have an example for using the "non-blocking" read_pulses in the Ciircuitpython adafruit_irremote library?
tulip sleet
@idle owl @slender iron do you understand why this travis build is failing? https://travis-ci.org/adafruit/Adafruit_CircuitPython_CircuitPlayground/builds/387798697?utm_source=github_status&utm_medium=notification
$ circuitpython-build-bundles --filename_prefix adafruit-circuitpython-circuitplayground --library_location .
Traceback (most recent call last):
...
File "/home/travis/virtualenv/python3.6.3/lib/python3.6/site-packages/circuitpython_build_tools/build.py", line 157, in library
shutil.copyfile(temp_file.name, output_file)
File "/home/travis/virtualenv/python3.6.3/lib/python3.6/shutil.py", line 121, in copyfile
with open(dst, 'wb') as fdst:
FileNotFoundError: [Errno 2] No such file or directory: 'build-adafruit-circuitpython-circuitplayground-py-0b681c3.zip/lib/adafruit_circuitplayground/express.py'
Is it because I removed __init__.py? I thought that file was superfluous if it's empty.
raven canopy
@marble hornet no need for attribution, but thank you for the consideration. @umbral dagger thanks for explaining the details. I was thinking along those lines, but couldn't explain it well enough.
stuck elbow
@tulip sleet wouldn't that be ImportError and not FileNotFountError?
tulip sleet
right, so I don't think that's the problem. It's like the expected zip file name is wrong
well maybe not zip file name, but it's very odd
onyx hinge
@tulip sleet in normal python anyway, __init__.py must be present to indicate that the directory is a Python package. I don't know about circuitpython specifically, though.
stuck elbow
@onyx hinge in "normal python" yes, in Python3 no
onyx hinge
apparently I'm not up to date @stuck elbow
stuck elbow
it's hard to keep up with what they change
prime flower
huh, MS acquired github today
prime flower
woah, that's a large sum of money...
ruby atlas
doubt it's all cash
stuck elbow
yeah, half is paid in juicero subscriptions
prime flower
@stuck elbow if we're going on their naming schemes you mean - Microsoft Sharepoint Code Revision R1 V24.4, Patch 3?
stuck elbow
could have been worse, could be Oracle...
actually, the bootloader we are using on the samd boards is by Microsoft too
they've been doing some good stuff with the embedded electronics
ruby atlas
Microsoft has been reinventing itself lately.
stuck elbow
I still have a hard time trusting them โ you know, most of the same people are still there in the management positions
it will take a generation or two
prime flower
didnt want to get into MS politics (there's a few employees on this discord), just express interest that the platform circuitpython is on got acquired
stuck elbow
sorry
not that we can't move the repos at a moment's notice
the automation would take some work to rebuild, though
onyx hinge
gitlab has CI but I've never used it
ruby atlas
works great.
stuck elbow
I wonder how many times this has to happen before people understand how to use a distributed version control system :)
onyx hinge
Yes, but the pull request model as implemented on github is decidedly centralized. and arguably it is the key feature of github that made it sensible for so many projects to centralize there.
slender iron
@tulip sleet that error should be reproducible locally. its something with the build script
idle owl
@solar whale Not that I know of.
solar whale
@idle owl thanks - it definitely behaves "differently" trying to make sense of it.
idle owl
It came out of a project I did - that's when we realised the IR code itself was blocking and nothing I did in my code could change that without changing the lib. I haven't looked at it since it changed though.
solar whale
If you enable it - simpletest and your example both don't work
its fine as long as blocking=True
idle owl
Good to know.
solar whale
I like the Neopixel control you did for the Ikea lamp demo - works nicely.
idle owl
Thank you ๐
marble hornet
thank you @umbral dagger
solar whale
I was thinking of a basic robot control for a CPX/Crickit using IR to control, but need it to be non blocking.
idle owl
Ah. So you're sorting through how to do that
solar whale
yes - I am making some progress, but was wondering how anyone has actually made the non-blocking work. Especially with the mini-remote
Having a lot of fun with the crickit
I also discovered that my monitor is a nasty source of IR noise! -- it went away when I disabled "energy saving" ๐ฆ
umbral dagger
@marble hornet Basically, yes. Only one object gets created, and put into each cell in the list.
marble hornet
thanks, it makes sense now
tidal kiln
anyone run cookiecutter lately? (for starting new cp lib)
File "./LICENSE", line 3, in template
jinja2.exceptions.TemplateSyntaxError: Encountered unknown tag 'now'.
File "./LICENSE", line 3
Copyright (c) {% now 'utc', '%Y' %} {{ cookiecutter.author }}{% if cookiecutter.company %} for {{ cookiecutter.company }}{% endif %}
@tidal kiln did you try updating cookiecutter? looks like its in jinja2-time dependency
tidal kiln
hmmm. actually. i just installed it. it wasn't on this box for some reason.
ah. hold on. i apt installed it...
@slender iron works with pip version. my bad. thanks.
slender iron
cool cool. no worries
wraith tiger
That looks like Postscript code.
tulip sleet
@slender iron what repo is the circuitpython-build-tools script in? I thought it was, uh, the circuitpython-build-tools repo, but apparently not.
tulip sleet
ah, ok, got it tnx
slender iron
<@&356864093652516868> meeting in a few minutes!
fluid helm
Hey @slender iron won't be able to make it tonight. Sorry ๐
slender iron
no worries @fluid helm
raven canopy
I'm late getting on the road, thinking to myself "I could just listen in..." Haha ๐
hollow ingot
Group hug and hugs to all the PyPI, Thonny and other maintainers (Punyforth, Espressif) for the work they've done and for patiently and capably handling issues.
raven canopy
Actually yeah. I'll be a fly on the wall. Force the kid to listen...
errant grail
Group hug and thanks to Dan and Scott for giving some suggestions for how to work with some memory limitations. Thank you!
wraith tiger
Just my standard: General kudos to all (group hug).
hollow ingot
Just knocking out a few dents.
hollow ingot
Built landscape and double-landscape Hackintosh and Windows 10 images for testing of cross-platform tools.
Built https://vgkits.org/blog/connecting-intro/ information - cross platform instructions for getting started (fairly generic to any ESP8266).
Built cross-platform vgkits-vanguard package to deliver command line tools to adopters of Vanguard, cutting the install steps and complexity opening up use of Circuitpython, Espruino, Adafruit_Blinka, and even Punyforth to a single command and a <20 second flashing routine.
See https://pypi.org/project/vgkits-vanguard/ for my first ever tool published to PyPI! It works too ๐
Started just trying to build out the pre-steps to https://vgkits.org/blog/connecting-with-miniterm/ and as the steps to get there became more and more insane had to solve it via building my own PyPI package.
errant grail
Besides pushing memory limits due to some large string arrays and trying to use OLED and TFT display libraries for text, been predominately heads-down revising some PCB designs. Reached my limit for hand-soldering and needed to make some adjustments. (Met my hand-soldering nemeses: SOT-723 and SOT23-5.) Still need to experiment with CPy memory management approaches but Iโm enjoying the learning process.
Will complete the CPy-based DSP-G1 MIDI synth this week. Also plan to send the next batch of four โRoverโ PCBs to OSH Park in a couple of days.
Off topic: my bandโs final gig is Saturday. A great 15-year run with my brothers from another mother. Mixed emotions.
... and bring back Carl and Jerry?
prime flower
(banned & muted)
slender iron
thanks!
hollow ingot
Sorry this is the best landing page for connecting to Vanguard (or any other CircuitPython/Micropython ESP8266 assuming you have the drivers) over USB using vgkits-vanguard https://vgkits.org/blog/vanguard-usb-howto/ welcome feedback and issues if this can be of use to anyone else. All the other links I shared are not so useful.
tulip sleet
@hollow ingot I took a peek at your pip3 guide. A lot of people end up doing sudo pip3 install because they get complaints about file permissions otherwise, but the sudo is usually a bad idea. pip3 install --user ... is the antidote to that, which might be good to mention if you think it would make sense.
umbral dagger
Awesome board. 2 wings on a metro
hollow ingot
Thanks @tulip sleet ! I forget to be explicit since on Debian/Ubuntu --user is the default.
idle owl
@ruby atlas https://oshpark.com/shared_projects/VHkoUs2Z
wraith tiger
I just did a bit of reading/research last week. This week I'm going to concentrate on working my way through the micro:bit micropython documentation and trying as many of the examples as I have the ability to try. I've found a few things to recommend documentation improvements for.
After getting through the documentation I'll check out bitio. Then I'll figure out what to do next.
ruby atlas
have to hop off
hollow ingot
Soon --user might be the pip default. Fingers crossed. However, there is debate about what the antonym should be. Debian went with --system, --global is now favoured I think. When that's in, they can consider --user as default cross-platform.
tulip sleet
@hollow ingot yeah I think it's not actually the default, or may not have been in 16.04.
wraith tiger
๐
errant grail
Thanks everyone!
umbral dagger
cool. later
prime flower
have an ultrawide - don't snap, too much space that still doesn't get used up
hollow ingot
@tulip sleet this is what the page should have read (including the --user flag) https://vgkits.org/blog/vanguard-tools-howto/ Sorry old page linked from step 3 was the wrong one (still pointed to the crazy complicated version which was eliminated by our PyPI package). This is the super simple one.
idle owl
@tulip sleet I'm looking at your PR on the CPX lib. That all works and doesn't create the touch object until you need it? I don't know of a way to test the backend stuff on CPX so I'm asking (I assume it functions for touch, I'm talking about whether it works for the reason we did all that nonsense in the first place to make it so it wouldn't init the pins and keep them in case you wanted to use touch with some external peripheral on one of the pins)
hollow ingot
@slender iron do you want to email me at cefn@cefn.com so I can CC you on the conversation with Phil on use of names?
BTW http://www.tannewt.org/ doesn't resolve to an IP address from here.
tulip sleet
@idle owl right, it still doesn't create the object until you call .touch_A2 or whatever. When you do that it look in ._touches(2), and sees whether there's a Pin object at the position or a TouchIn object. If there's a Pin object, it replaces it with a TouchIn(that pin) object.
idle owl
Ok keen.
tulip sleet
i'm working on why travis is failing. It is because there's no __init__.py, but there's something weird about the build.py script that's causing the error.
idle owl
Oh ok right
slender iron
@hollow ingot I'm scott@adafruit.com
idle owl
I can't merge it yet anyway, forgot travis failed.
slender iron
@hollow ingot and I should add www. support. it works without it
slender iron
no need to blame them
hollow ingot
Teehee
idle owl
@tulip sleet Can we also remove the two LIS3DH try/except blocks at this point? The ones that say you need to be running 2.2.0+. I feel like that should be a given at this point. Even the ones that shipped with it shipped higher.
I also had another question about how we're checking compatibility and whether one way is more efficient than the other.
(I told myself to write this stuff down, and I didn't, nearly forgot.)
tulip sleet
is that for ustruct vs struct and collections vs ucollections??
idle owl
Neither, it was because there was no shake functionality in the older version.
Or tap.
Had nothing to do with struct or collections.
tulip sleet
oh, oh, you mean in Express. I was looking in the LIS3DH library.
idle owl
Oh, sorry. Yes.
I meant they were in LIS3DH blocks in the Express class, sorry
That would get rid of 8 lines of code and however many characters they have.
ruby atlas
wow the price diff for 3 vs 10 boards isn't very big at oshpark
idle owl
Oh and I think we can remove it from detect_taps too, I think that's why there's a try/except in there as well.
tulip sleet
@idle owl I think that's fine. It's still an error, and we can cover any issues in the forum. It will save yet more space. I'll make another commit with that.
idle owl
@ruby atlas I think they price by area, so it all depends on what you're requesting.
@tulip sleet Ok good.
ruby atlas
oh, yeah, looks like a medium run is smaller than that board's area x 10 ๐
idle owl
@tulip sleet II have one more question too, but it's a little more complicated to type I think
idle owl
We can. So did I, heh
I think maybe what I just suggested fixes my next question. But it's worth having the answer anyway.
I'll brb. Then we can chat.
idle owl
Ok
slender iron
<@&356864093652516868> Here is the recording from today's meeting: https://youtu.be/QTwU0tLXZm4
idle owl
@tulip sleet Ooh. So I added *tone() code to the sound_level code and it works! (While the code is still running anyway). I can use loud_sound() to trigger a tone and there's no issues. So the PDMIn fix must have resolved that.
idle owl
Correct, it fails simply printing the sound_level to the serial console.
tulip sleet
๐ฆ
idle owl
It eventually hangs with the extra code added too, but it's hanging the same way without anything else in it, so it's not related to running tone and mic at the same time. It's still the same hang.
(So cool though, I can blow on it and it'll play a tone for the duration of the breath and stop when I stop.)
I keep jumping into the REPL to stop it from getting to the point of hanging.
Have to jump into the REPL to edit the code too or it'll hang while trying to edit the code and then the code doesn't save.
marble hornet
@tulip sleet and @idle owl may i ask why the 'u's as added to some of the name. and do you know where uheap went in 2.3.1?
idle owl
It's when it's a carryover from MicroPython - u is shorthand for mu, which is the symbol for micro.
tulip sleet
The us came from MicroPython's versions of these modules. They made u versions for a number of standard regular Python modules. We are trying to go back to the original names, and make our versions be subsets of the original functionality, so you can move the code over to regular Python easily. The MicroPython versions were not necessarily proper subsets.
uheap is a debug module -- I don't think it's turned on in general. You have to enable it in the build.
marble hornet
ah, thanks!
tulip sleet
and kattni is pointing out the origin of why they choose "u" as the prefix
marble hornet
ฮผ?
idle owl
Correct
tulip sleet
ya
marble hornet
thanks
prime flower
or ฮผCenter (Microcenter) if you're retail-inclined
tulip sleet
ฮผPython
marble hornet
is there a way to check the amount of unfilled memory?
wraith tiger
...and ฮผflash, though it's written uflash, it's pronounced "microflash."
marble hornet
thank you ( i can't wait for my samd51 pcb comes in!!)
@idle owl @tulip sleet thank you
tulip sleet
@idle owl is this also 2.2.0 protection?
try:
if value == 1:
self._lis3dh.set_tap(value, 90, time_limit=4, time_latency=50, time_window=255)
if value == 2:
self._lis3dh.set_tap(value, 60, time_limit=10, time_latency=50, time_window=255)
except AttributeError:
pass
Yes.
tulip sleet
got it will remove the try-execpt wrapper
tulip sleet
removing the try's saves another 264 bytes, for a total of 628
tulip sleet
tnx - worth it
tidal kiln
@slender iron in process of setting up new repo for lib work..got to step on setting up readthedoc service for the repo on github -> there's a warning banner saying "Services" are being deprecated. (your guide shows this process)
idle owl
@tulip sleet Ok tested. Merging. Do you want to do the release?
tulip sleet
sure
surreal saffron
Any feedback on https://github.com/adafruit/circuitpython/pull/891 ? 
rigid path
Is there a guide out there for getting an M0 Adalogger Feather setup for Circuit Python use? I found a start guide but I only see Arduino info, (https://learn.adafruit.com/adafruit-feather-m0-adalogger/) or am I just not reading this correctly?
tidal kiln
@rigid path check this one: https://learn.adafruit.com/data-logging-with-feather-and-circuitpython/overview
marble hornet
are ```"""
many line comments"""
uses the logger feather wing, but general idea would be the same
@marble hornet i think so
rigid path
@tidal kiln -Thank you, I will check it out.
tidal kiln
yes. multiline string literals. but when used as comments, they aren't being assigned to anything.
>>> g = "hello"
>>> g
'hello'
>>> g = """hello
... world"""
>>> g
'hello\nworld'
>>> "hello"
'hello'
>>> """hello
... world"""
'hello\nworld'
>>>
but when used in certain locations, they are considered docstrings
slender iron
@tidal kiln does it have a link to more info?
surreal saffron
What does this do in a repl?
class Foo: 'docstring'
Foo.doc
What does this do in a repl?
class Foo: 'docstring'
Foo.doc
... what's code mode? ๐
test
idle owl
@surreal saffron It's three backticks on either side for a code block or one on either side for inline code. The backtick is in the upper left hand corner of a US keyboard.
surreal saffron
Thanks.
tidal kiln
@slender iron yep. this looks like the one with the deets: https://developer.github.com/changes/2018-04-25-github-services-deprecation/
surreal saffron
As I was saying...
What does this do in a repl?
class Foo: 'docstring'
Foo.__doc__
That'll tell you if docstrings are kept around.
reef seal
@slender iron Could you give me more info on your comment: https://github.com/adafruit/CircuitPython_Community_Bundle/pull/9
Not sure what you're asking for
slender iron
@tidal kiln yup, I've been doing plain webhooks and the guide needs an update
@reef seal the commit the submodule includes should match the latest release from the sub-repo
idle owl
Sphinx is failing on the first line and I'm not sure why. It's obviously something not entered right but I'm looking at other libs and they all start that way. By first line I mean 23. First line of the docstring. https://github.com/adafruit/Adafruit_CircuitPython_MatrixKeypad/blob/master/adafruit_matrixkeypad.py
slender iron
@idle owl my guess is capitalization
idle owl
@slender iron Thank you I will try that.
slender iron
@reef seal yup, plus updating the PR to point to that commit
reef seal
If I tag the existing latest commit that's ok? Or I need to generate a new commit also?
As long as it's pointing to the latest release I think that's what you're asking for
slender iron
@reef seal existing commit is just fine. its the release that matters
our adabot process will automatically try and update it to the latest release
tidal kiln
@idle owl looks like it passed?
idle owl
@slender iron THANK YOU.
reef seal
cool
slender iron
๐
tidal kiln
oh. it was the previous commit?
idle owl
I put in another commit on Scott's suggestion and it fixed it.
@slender iron For adding it to RTD, who am I supposed to be logged into RTD as? Because I'm in as me and it doesn't look right, there's not the list I think I'm supposed to be seeing.
slender iron
I do it as myself
idle owl
hmm ok
idle owl
I have that in my notes.
slender iron
the other repos should have it
idle owl
@slender iron I'm adding the webhook, and I have in my notes to "select individual events" and then I have "create" and "push". I don't see one that is create, there's a bunch with create in the description of what they are, but I don't think that's right. What do I select when I'm adding the webhook URL to GitHub?
slender iron
I think for releases
idle owl
There is a "releases" check box
tidal kiln
@idle owl ha! i just went through this
idle owl
Should I add "Releases"?
tidal kiln
which side are you setting up? github or rtd?
idle owl
GitHub
tidal kiln
are you under repo -> Settings -> Webhooks?
idle owl
Yes, already clicked add new, pasted the URL, now I'm choosing the individual events on which to trigger the webhook
tidal kiln
do you have this link in your notes?
https://docs.readthedocs.io/en/latest/webhooks.html#github
idle owl
No.
So the answer is no, push is fine. And I don't need to select individual events, I can leave it on "just the push event" apparently?
tidal kiln
think so give it a try
i'm actually testing this right now for a repo i just created...
idle owl
Seems to work so far. Adding to CircuitPython now.
@tidal kiln The alias, does it have underscores in it, like adafruit_circuitpython_libname or is it without underscores
tidal kiln
alias where?
idle owl
When adding a subproject to CircuitPython
oh nm
I looked at another one.
it's just libname
tidal kiln
is this a submodule?
idle owl
it's a library. The subproject list in CircuitPython is the big list I remember seeing. It isn't that I was logged into RTD under the wrong login, it's that I hadn't gotten to that step yet.
That's what gets it to cp.rtd.io/projects/libraryname
@tidal kiln I requested your review on adding the lib to the bundle.
hmm.
Wait really? I thought it didn't care if there were multiple .py files in the examples folder. I thought it was only an issue in /
/home/travis/build/adafruit/Adafruit_CircuitPython_Bundle/libraries/matrixkeypad/examples
Multiple top level py files not allowed. Please put them in a package or combine them into a single file.```Dinner is here. I'll be back in a bit.
tidal kiln
@idle owl added comments in PR for when you get back
marble hornet
@errant grail i'm also working on text libraries for circuit python, wanna chat. i'm curious how you are approaching it. mind chatting?
and i know it is late but:
tulip sleet
@idle owl
[submodule "libraries/drivers/stmpe610"]
path = libraries/drivers/stmpe610 path = libraries/drivers/stmpe610
url = https://github.com/adafruit/Adafruit_CircuitPython_STMPE610.git url = https://github.com/adafruit/Adafruit_CircuitPython_STMPE610.git
+[submodule "libraries/matrixkeypad"]
+ path = libraries/matrixkeypad
+ url = https://github.com/adafruit/Adafruit_CircuitPython_MatrixKeypad
should libraries/drivers/matrixkeypad or libraries/helpers/matrixkeypad, I think
marble hornet
#hug_report @raven canopy , @tulip sleet , @solar whale, @slender iron , @stuck elbow, and all the other names that are escaping me. Thank you for answering my prodigious amount of questions. iv'e been able to make leaps not strides in my project. couldn't have done this without you.
tulip sleet
happy to help!
tulip sleet
ga
marble hornet
github
ah!
so;
when i'm making alterations to a cloned repo, and have been doing it for a while, how can i get the most recent updates to the master w/out overwriting my previous local edits ?
did i say that well enough?
tulip sleet
sure, so you're working on code in a repo, and meanwhile it's changing out from under you on master. So make a branch and commit your changes to the branch, or "stash them"
tidal kiln
@marble hornet i think you're asking this: https://help.github.com/articles/syncing-a-fork/ maybe?
tulip sleet
cater has the info, then
marble hornet
tnx
tulip sleet
work in a branch, and you can keep resyncing master, and merge or rebase the changes from master into your branch
slender iron
@surreal saffron reviewing your PR now. Thanks for the patience!
tidal kiln
@idle owl ๐ค look now?
tulip sleet
yeah, the comments don't show until you submit a review
idle owl
Yah they're there now.
I followed the instructions in the "adding a library" in the readme for the bundle.
I don't know how to add the path, manually edit the file after doing the git submodule add ?
tidal kiln
forgot to click the submit review button
idle owl
oh wait.
tidal kiln
i think i've just manually edited the .gitmodules file in my fork, then PR'd that back
tulip sleet
Try git mv matrixkeypad drivers. I think they finally made that work in git. Try that and then check .gitmodules
tulip sleet
correct the bundle README?
idle owl
I'm not sure how to update it. I should have looked at the structure of the rest of the libs, I guess. (How as in what to update it to, not how as in the actual updating of it.)
tulip sleet
It says:
Determine the best location within libraries for the new library and then run:
So maybe just say "Determine the best location within libraries (libraries/drivers/ or libraries/helpers/) for the new library and then run:"
idle owl
Passed!
tidal kiln
passed
idle owl
@tidal kiln Do you have merge permissions on that repo?
tulip sleet
it's not at all
idle owl
@tulip sleet Should I close the record/playback issue since it's fixed in master? The hanging isn't specific to that.
@tidal kiln Oh right I forgot to check that since it showed both of your review points as fixed when I added the different folder structure.
tulip sleet
i think it may be the same bug, or maybe not. You can close it in favor of keeping #879 open.
idle owl
ok.
tulip sleet
You can say either this bug is fixed or it's another manifestion of #879.
idle owl
Yah on it
idle owl
@tidal kiln I can add it and push again
Actually there are others without it.
Found them scrolling down to get to the bottom.
AMG88xx and Seesaw
tidal kiln
are there? meh. then i'll get over it. don't worry about it.
tidal kiln
want me to merge?
idle owl
Yes please
tidal kiln
done
tulip sleet
@slender iron are you willing to take look at #879 a bit? I wrote some more notes. Perhaps some fresh eyes would help. I'm happy to look at it some more but not tonight. tnx.
slender iron
I have a bit of time now @tulip sleet
idle owl
@tulip sleet What's the current status of ESP32? Still "we're not developing for it, but we'll pull it from MicroPython when dev there for it is complete. No ETA." ?
In terms of CircuitPython.
tulip sleet
right
idle owl
Thank you
@tidal kiln Do the CP boards need to be placed in the bootloader to load an Arduino Sketch?
tidal kiln
can't remember.
idle owl
Oh
tidal kiln
but bootloader definitely works
tidal kiln
easy enough to test
idle owl
If you have Arduino installed, yes.
I never went through getting it all setup again after my other machine was stolen.
tidal kiln
i just do it so fast and often i'm not sure
idle owl
Ah
tidal kiln
so scenario is you've got a board running CP and want to program Arduino?
idle owl
Yes. The actual scenario is "How do I switch between CircuitPython and Arduino", but that is the part I'm answering right now is CP to Arduino
idle owl
My answer was "If you're currently running CircuitPython, double-tap the reset button on the board to get the board into the bootloader, and then load your Arduino sketch."
tidal kiln
didn't need to reset
tidal kiln
seems like. testing it various ways.
tidal kiln
i'm pretty sure it's meant to do that, but there may be some weird edge cases that may require a double reset.