It wasn't added to travis correctly.
#circuitpython-dev
raven canopy
.. image:: https://travis-ci.org/adafruit/Adafruit_CircuitPython_L3GD20.svg?branch=master
:target: https://travis-ci.org/adafruit/Adafruit_CircuitPython_L3GD20
:alt: Build Status
should be this ^^^
raven canopy
Travis failed on the deployment invalid option "--api_key=".
https://travis-ci.org/adafruit/Adafruit_CircuitPython_L3GD20
idle owl
Correct
raven canopy
the RTD part...is beyond my realm. other than it needs to be added to the circuitpython drivers list...
fading solstice
I will fix that, but the next thing is to get it into the Adafruit CircuitPython RTD.
raven canopy
needs to be added here: https://github.com/adafruit/circuitpython/blob/master/docs/drivers.rst
I think adabot handles the RTD build. that correct, @idle owl?
idle owl
Not exactly...
raven canopy
hehe. and that is why i still don't understand it fully. ๐
idle owl
We're talking about adding the lib to RTD so it has an RTD page, correct?
raven canopy
yes. it doesn't have one built, best i can tell.
idle owl
Yes I have half a page of notes on this. It's a series of steps done on RTD.
fading solstice
is that something i should do (can do)?
idle owl
Let's find out!
@slender iron Is adding a lib to RTD something that @fading solstice has permissions for?
raven canopy
I'll hold off on approve-n-merge for the badge fix. I don't think there is anything to do in the lib repo for RTD, but just in case...
idle owl
Only getting the URL once it's done, but waiting is fine.
fading solstice
after i push the fix to my repo version, i am getting a passing from the builfd badge. but stilll unknown from the docs
idle owl
I think because it's not been added to RTD.
raven canopy
it isn't listed in the sub projects; just need to find out if adabot will/should do it after adding to drivers.rst, or if it requires addition by hand.
peak thicket
Hi,
manic glacierBOT
peak thicket
anyone flashed the latest circuitpython on a nrf52? any good results?
solar whale
@peak thicket yes - it works on the feather52832
slender iron
just ordered an odroid-go ๐
ruby atlas
soo many things i want to order
@slender iron I take it 3.x is the RC branch for 3.0, and that my PR should remain against master?
raven canopy
vagrant ssh does not belong in this window... ๐
manic glacierBOT
solar whale
@tulip sleet I did not realize there had been a 3.x branch -- I applied the itsybitsy m4 fix for Pins D3/D4 to master - does it also need to get applied to 3.x or will that happen at the next 3.x release.
The ESP8266 boot sequence is making my head hurt ๐จ
tulip sleet
@solar whale I'll look at the ESP8266 problems tomorrow. Don't worry too much for now. I'll cherry-pick the D3/D4 fix over to 3.x.
solar whale
@tulip sleet thanks - I mades ome progress - posted to the issue, but I'm stuck now. Glad to leave it to you.! Good luck.
raven canopy
i was reading what you were finding @solar whale. it seems to me that the esp was never expected to have the FS locked. otherwise, ampy wouldn't have worked anyway. right?
tulip sleet
any esp8266 fix will go into 3.x and I'll merge it back to master
solar whale
@raven canopy I think so, but I can't figure out where it gets locked.
raven canopy
i think it's the updated boot_out stuff. unless i misunderstand the changes you made there @tulip sleet?
solar whale
but also one of the recent commits caused FS to appear corrupted.
but taht is in main.py at root/ esp8266 does not use that .
raven canopy
right, based on the bdev not being empty, in inisetup.py.
raven canopy
oh yeah...argh! too many thoughts and material at once. ๐
solar whale
but the bdev seems to be the problem - or at least it is related. I think the FS is never getting mounted since it fails in _boot.py
raven canopy
think we're hitting this: https://github.com/adafruit/circuitpython/blob/master/ports/esp8266/modules/flashbdev.py#L31? it would explain the NoneType error.
but why are we now hitting that is the caper...
solar whale
Thats's where I have been looking -- but hard to see why it would fail
raven canopy
if only we had some debug tools... ๐
tulip sleet
aha, I know the problem, I think, thanks for the pointer to that file. readblocks() and writeblocks() need to return a status. I am now looking for a return code. Formerly it didn't check for errors. We need to add return 0 to the end of each of those rooutines
there was a TODO lemme find that
extmod/vfs_fat_diskio.c diskread():
mp_call_method_n_kw(2, 0, vfs->readblocks);
// TODO handle error return
this used to be true, so I handled the error return, but it turns out a number of file system impls were not bothering to return any status. I had to fix the tests where a toy filesystem was not, for instance.
Is it easy for you to add the two return 0 lines and test? That is probably the problem.
solar whale
I'll give it a try
raven canopy
readies Captain Planet ring
With our powers combined... ๐ ๐
this unsigned math -> MP_NEW_SMALL_INT thing is driving me batty. I don't get it... I see it used successfully elsewhere, my sign bits aren't changing (that i can tell), and the 1 deep ring buffer is functioning correctly so I shouldn't be over/underflowing.
solar whale
first try did not work, but I had done some reverts -- unrevrting so only change is to flashbdev.py
nope back to corrupted FS when I put the removed commits back. ```erryneedell@Ubuntu-Macmini:~/projects/adafruit_github/circuitpython_master/ports/esp8266$ git diff
diff --git a/ports/esp8266/modules/flashbdev.py b/ports/esp8266/modules/flashbdev.py
index 40ba655..da4194b 100644
--- a/ports/esp8266/modules/flashbdev.py
+++ b/ports/esp8266/modules/flashbdev.py
@@ -13,12 +13,14 @@ class FlashBdev:
def readblocks(self, n, buf):
#print("readblocks(%s, %x(%d))" % (n, id(buf), len(buf)))
esp.flash_read((n + self.START_SEC) * self.SEC_SIZE, buf)
-
return 0def writeblocks(self, n, buf):
#print("writeblocks(%s, %x(%d))" % (n, id(buf), len(buf)))
#assert len(buf) <= self.SEC_SIZE, len(buf)
esp.flash_erase(n + self.START_SEC)
esp.flash_write((n + self.START_SEC) * self.SEC_SIZE, buf) -
return 0def ioctl(self, op, arg):
#print("ioctl(%d, %r)" % (op, arg))
it's also possible it's forbidding '/' to be mounted on '/', though it's supposed to handle that. All right, enough for tonight.
I think there may be multiple issues.
solar whale
agreed -- good night -- and good luck !!
raven canopy
night all!
tulip sleet
so says Edward R. Murrow...
raven canopy
admits: had to look him up. but i know after the catchphrase. ๐
ok. time to turn the brain off... ๐คค
timber mango
For @raven canopy - 2's complement (signed ints). This is the first draft (I rarely do a second) hehe.
indigo wedge
is there a port of https://github.com/adafruit/Adafruit_TouchScreen for CP? I know it's not much code but want to make sure someone didn't already do it before I try myself ๐
manic glacierBOT
solar whale
@indigo wedge I did craate a driver for the STMPE610 touchscreen if that is of any help https://github.com/adafruit/Adafruit_CircuitPython_STMPE610
but it looks like you are going straight to the 4 wires.
indigo wedge
Yes, I wanted a 4-wire one, oh well, no worries
tulip sleet
@slender iron Did not mean to merge this PR without your review: something odd happened when I thought I was pushing to my local repo, and the PR got merged. But it doesn't really matter. This was just cherry-picking a fix from master that's known to work. https://github.com/adafruit/circuitpython/pull/952
tulip sleet
@solar whale are you available to test an esp8266 fix?
marble hornet
does anyone have a suggested smd button with nice click (tactile or not)? the one i choose seems to be crappy. many thanks either way.
solar whale
@tulip sleet not until about 4 PM
marble hornet
it kinda shorts out / messes with the reset and ground rail
tulip sleet
I have a fix for the esp8266 problem. The fix allows '/' to be mounted on '/'. with the fix, ampy works fine and no errors on startup.
solar whale
@marble hornet these work well https://www.mouser.com/ProductDetail/667-EVQ-Q2503W
tulip sleet
The esp8266 startup appears to mount a '/' twice instead of just once.
marble hornet
thanks!
tulip sleet
once is in code that isn't error-checked, and once is in the startup .py code, which I added a check for.
I'll submit a PR
solar whale
@tulip sleet nice - I'll be happy to try it but is sounds like it is working well. Thanks!
tulip sleet
also I reworked the error checking so the return 0s aren't necessary: a filesystem impl can return None or 0 on success. Any other return is an error. Not important to you but there might be other ports where there's no status return.
k, tnx, ttyl
tulip sleet
@slender iron increased GitHub branch protection for 3.x branch - made it same as 2.x.
indigo wedge
Hey @tulip sleet could you have a look at my nRF patch when you get some time, I know it's rather large but if you go commit by commit then it's pretty clean I think ๐ Thanks!
tulip sleet
@indigo wedge sure, one or more of us will. I have not done any nRF work yet so Scott, hathach and I will work out who reviews what.
wraith tiger
Has anyone asked the PSF to list CircuitPython on https://www.python.org/download/alternatives/?
plucky flint
@wraith tiger you can submit an issue to the PSF's web team. I believe the repository is on GitHub (but I don't have it to hand). If nothing comes of it, I know one of the maintainers and can give him a kick if there needs to be movement.
slender iron
@tulip sleet Thanks for the fixes. No worries on my end about the merged PR. I've been heads down on this gaming stuff though I want to get caught up today
@ruby atlas yup, you should target master
timber mango
I am using several adafruit feather boards in ubuntu and all works fine when I am running as the logged in user, with the CIRCUITPY device mounted at /media/<user>/CIRCUITPY. However, I need to copy files to the device to run code from inside jenkins. So - can I modify the mount point in fstab so the device is available to eveyone, or can I modify the group permissions for the device mounted at /media/<user>/CIRCUITPY so jenkins can write the code.py file to it ?
slender iron
@timber mango I imagine it would require a change to udev rules. I don't know the details though
probably the same as other USB drives
timber mango
tannewt - I have the adafruit udev rules already setup - the issue is the way ubuntu now automounts all usb devices to /media/<user> and the permissions that are set by default. I am having trouble finding a decent reference online that shows how to change the permissions for the group. I 'fixed' a similar issue for the second hard drive in my linux system by having it mounted via fstab and not in the media folder. I will keep looking as it should be a straight up permissions edited (but I have not found the correct way to do that yet - the joys of open source)
analog folio
could I get some pointers on a Python issue in having here?
slender iron
@analog folio ask away
anyone know where the nrf52 install instructions are? https://forums.adafruit.com/viewtopic.php?f=60&t=137289
solar whale
just posted it to the forum
slender iron
haha, I did too
solar whale
good - mine got blocked by yours -- all good
tidal kiln
@analog folio responded in #help-with-projects
wraith tiger
So, if I were to request that CircuitPython be listed on the PSF's "Alternative Python Implementations" page, what one sentence blurb would we want to use and what URL should it point to? I'm guessing https://learn.adafruit.com/welcome-to-circuitpython would be the right URL.
solar whale
@tulip sleet @raven canopy can you point me to some information, or give a brief explanation of waht this funtion is doing MP_OBJ_SMALL_INT_VALUE(ret) it seems to be a hot topic these days...
solar whale
it look like it just divides by 4
tulip sleet
@solar whale it is a macro that converts from object representation to a regular integer
#define MP_OBJ_SMALL_INT_VALUE(o) (((mp_int_t)(o)) >> 1)
solar whale
oh - I picekd the wrong one to look at
tulip sleet
32-bit values can be object pointers, interned string pointers, ints, or floats. The lower 1 or 2 bits indicate the kind of value.
if the lowest bit is set, the upper 31 bits are an integer
solar whale
static inline bool MP_OBJ_IS_SMALL_INT(mp_const_obj_t o)
{ return ((((mp_int_t)(o)) & 1) != 0); }
#define MP_OBJ_SMALL_INT_VALUE(o) (((mp_int_t)(o)) >> 1)
OK thanks --- fog is lifting.
How do you know if it is REPR_A/B/C/D ?
tulip sleet
If the lower two bits are such and such, it's a float, etc.
repr choice is in mpconfigport.h, probably
yes, ok, it's REPR_C, chosen in ports/atmel-samd/mpconfigport.h
solar whale
ah - thanks
also in ports/esp8266/mpconfigport.h ๐
and I thought I could read C programs ....
tulip sleet
I use ag a lot.
manic glacierBOT
raven canopy
@solar whale thanks for joining me in MP_OBJ_NEW_SMALL_INT land! ๐ i'm about to try mp_obj_new_int_from_uint() to try and confirm that i am somehow sending an incompatible uint...
slender iron
@fading solstice did you get the RTD stuff sorted?
raven canopy
@slender iron i don't think so. we left off at "how does the RTD build get triggered/created"...
slender iron
the RTD project is created manually (usually by me). the builds get triggered through a github webhook (also set up manually)
raven canopy
ok. so those need to happen. along with drivers.rst being updated.
slender iron
yup
raven canopy
i'll approve and merge the Travis badge fix PR, if it doesn't affect any of that.
ohh. nevermind. you did that. 7 minutes ago... ๐
slender iron
๐
timber mango
Is there a good visualization tool for git that can run on the local Linux machine I'm editing on? I want to see the diffs before I push commits to a branch.
This way I keep the edit history a bit more tidy, when other folks look at the code on github.
raven canopy
i'm thinking about tonym01's uf2 update lockup issue. kind of easy to blame the PC, but i'm wondering if it could be an older version of the bootloader that is giving them fits. this is an issue that is "close", but the symptoms don't exactly match: https://github.com/Microsoft/uf2-samdx1/issues/30
also, having them update would kind of erase the "evidence" they currently have. is there a way we can get a dump of the installed uf2?
tulip sleet
@timber mango I use gitk. There's also git-gui.
timber mango
The following NEW packages will be installed: git-gui gitk libtcl8.6 libtk8.6 tcl tcl8.6 tk tk8.6 0 upgraded, 8 newly installed, 0 to remove and 0 not upgraded. Need to get 3,415 kB of archives. After this operation, 9,798 kB of additional disk space will be used.
@tulip sleet thanks! I found those while waiting for a reply here, so they're about to be installed. Good to know you like to use them!
https://git-scm.com/download/gui/linux is what put me onto them.
tulip sleet
@raven canopy Since it locks up based on the CPy version, I'm not sure how the UF2 bootloader would be implicated. I just got my Win10 machine back after my wife borrowed it this afternoon, so I'll try this too.
@timber mango another harder q is what tool to use if there's a merge conflict (you probably won't encounter that), il.e. what git mergetool uses. I use kdiff3.
raven canopy
@tulip sleet fair point. the main reason i was thinking bootloader is because in boot mode it is the one setting the USB descriptors. I am tempted to help them with the microsoft USB diagnostic tool that I used with serial paste mode. but it isn't easy to read/find info...
tulip sleet
When tonym01 says the mouse and keyboard are unresponsive, I am wondering if it's the HID descriptors that 3.0 presents, which are more complicated than the 2.3.1 descriptors.
3.0 has mouse, keyboard, gamepad, and consumer control all rolled up. Or, it could be a disk mounting issue. I've seen third-party software get confused
raven canopy
yeah. the "unlocks/locks on reset" would support that. the original lockup description doesn't clearly state if it locks after the copy is complete, or before.
tulip sleet
ok, I can't duplicate. I'll have them erase the filesystem and see if it's some peculiarity there.
ruby atlas
@tulip sleet "mouse and keyboard are unresponsive" I get that when I crash CP.
Or am running the JLink
I'm not sure if it's HID or Mass Storage causing it. Mass storage usually disconnects shortly after that
tulip sleet
does it recover after a little while or do you have to unplug the board?
ruby atlas
it does recover (this is a Mac), but stalls intermittently until I unplug the board.
tulip sleet
The stalling may be that's it's upset about the mass storage being disconnected. @slender iron have you seen anything like this on your Mac?
ruby atlas
@tulip sleet i suppose I should probably watch my logs sometime.
When working with the JLink next, since I can simulate a hang that way.
tulip sleet
yes, whatever the equivalent of syslog is would be interesting
ruby atlas
Console ๐
Turns out it's still syslog: $ ps ax | grep syslog 50 ?? Ss 0:39.08 /usr/sbin/syslogd
tulip sleet
tail -f /var/log/syslog ?
ruby atlas
I would have expected Apple to replace that with something custom by now.
tulip sleet
probably somewhere else
ruby atlas
tail -f /var/log/system.log and a few others.
but Console app is better for a more overall view of all the logs, as it's likely not in the system.log
tulip sleet
I am not a heavy Mac user - I have an old one for cross-platform testing only.
ruby atlas
Ah.
slender iron
I haven't seen the issue @tulip sleet
ruby atlas
Let me go read the issue.
Wow that's kinda ugly, way worse than the brief USB HID issues I feel when using JLink debugging (or crash CP) - 1-2 second HID device hang before the drive is ejected.
It's a bit worse on my desktop mac than my 2016 macbook pro.
my desktop mac isn't using official apple hardware.
timber mango
@tulip sleet how do you scroll the lower left window in gitk?
It's showing a DIFF.
The keyboard arrow keypresses do not scroll this window (unshifted, anyway).
Instead, they select the previous or next commit.
(The mouse scrollbar there does scroll the DIFF, but is difficult for me to operate.)
tulip sleet
it scrolls automatically as you scroll the lower right window, or select files in there. It's not that great: it's just all the file diffs concatenated together
oh, you mean if the diff is taller than the window?
timber mango
I use huge fonts so the diff is never going to fit inside the window -- I have to scroll it.
tulip sleet
i just move my mouse into the lower left subwindow and use the scroll wheel, and it is scrolling
timber mango
Okay got that. I can do an analogous thing with my trackpad. Thank you.
tulip sleet
or drag that subwindow's scroll bar
timber mango
Scroll bars require better vision to be useful. They are strongly visually-biased widgets. ;)
The Help has keybindings; This is Very Linuxy.
u is up and d is down.
The increment is 18 lines and doesn't seem to be configurable, but maybe it is.
It also binds Ctrl Minus and Ctrl Plus to font decrease/increase.
Seems to function somewhat like a $PAGER. Not too bad.
Definitely ahead of where I was, two hours ago.
frosty bison
I have a question about mpy-cross. Is the original micropython mpy-cross output compatible with circuitpython? Or do I have to use the circuitpython fork of it? There's a binary available for micropython, but not circuitpython from what I can find. And no instructions provided to build circuitpython on Windows 7 which would probably be a pain anyway.
solar whale
@frosty bison what version of Circuitpython are you using -- there is a different version of mpy-cross for 2.x from 3.x
frosty bison
@solar whale adafruit-circuitpython-trinket_m0-2.3.1.uf2 is what I'm running on my Trinket M0
frosty bison
Perfect, I think I've got it from there, thanks for the quick reply
solar whale
not sure if it is windows 10 or 7 or possibly both...
@slender iron @tulip sleet I don't see a CP3.0 mpy-cross with the 3.0 releases - am I just not seeing it?
frosty bison
the mpy-cross at https://github.com/adafruit/circuitpython/releases/tag/2.3.1 works great so I'm good to go
slender iron
@solar whale its on older ones but I'm bad at copying it over
ruby atlas
should we be renaming mpy-cross for 3.x builds to mpy-cross-3.0 and 3.1 and so on?
I have run into a few cases where I needed more than one mpy-cross version.
wraith tiger
@timber mango I'm using Git Cola (https://git-cola.github.io/).
solar whale
so far - I have only found a windows version back with 3.0 alpha 6 -- I guess it has not been a problem - yet...
timber mango
This gitk visualization of the diff gives me a chance to almost be sober (hah) in how I approach this (I have sweeping changes to make in my code base, which I inherited and disagreed with the old method of representing the code in very looong files).
(insert a few words and joining concepts there, some'eres)
manic glacierBOT
raven canopy
@slender iron i've meant to ask you this before, but can you explain (or point me to) why "TC4-TC7 can only have 100MHz inputs", per this note: https://github.com/adafruit/samd-peripherals/blob/master/samd/samd51/timers.c#L122
slender iron
take a look in the SAMD51 datasheet under electrical characteristics
its got a list of input GCLK limits for peripherals
raven canopy
kk. thanks!
tulip sleet
@ruby atlas @solar whale any 3.0 mpy-cross will work. I need to gather them together and add them to the release.
ruby atlas
I just discovered they still make something like my childhood electronics learning toy https://www.amazon.ca/Elenco-Electronic-Playground-Learning-Center/dp/B0035XSZDI
ruby lake
is there a preferred way to use an equate or define for a value without comitting it to a variable?
something like #define noteD 1 in C code
stuck elbow
you can do _variable = const(5), but it will only be available in the one file in which you put it
tulip sleet
@ruby lake you need to from micropython import const first. Details here: http://docs.micropython.org/en/latest/pyboard/library/micropython.html#micropython.const and http://docs.micropython.org/en/latest/pyboard/reference/speed_python.html#the-const-declaration. The rest of the second page is also interesting.
onyx hinge
#954 is sure a weird issue
stuck elbow
windows ๐คท
tulip sleet
@slender iron q re what's created on imports: I have a .py that's just a couple dozen FOO = const(123) sort of lines. The .mpy is only about 600 bytes. When imported either as import foo or from foo import *, about 3k bytes gets used up, which seems kinds of excessive to me. What's your intuition here? Is it the dict and the pointers in the dict? How sparse is the dict? Perhaps I should use your heap analysis script. I wonder if there's some way to shrink that down significantly.
hmm, if I import just a single name: from foo import FOO, it still uses up 3k bytes, so it's not just the global name dict
stuck elbow
the variable name has to start with _ to save memory
otherwise it's added into the dict anyways
because it has to be available from the outside
river quest
thank you everyone for a great year here together! https://blog.adafruit.com/2018/06/22/celebrating-a-year-together-as-a-community-on-discord-discordapp-adafruit-adafruit/
wraith tiger
@ruby atlas I had something like that, but I think mine was from Radio Shack.
slender iron
@tulip sleet It could be qstr pool related. probably good to get you running the heap script. it could be a long lived leak as well causing duplicates
in your example FOO still refers to its namespace so all of the other names are still in its dictionary
prime flower
@wraith tiger https://blog.adafruit.com/2018/06/22/when-ibm-hired-musicians-as-programmers-retrocomputing/
tulip sleet
tnx to you and deshipu for that point. I am going to try one-name .py files as well, which seem to use up too much space. Will come up with an example.
wraith tiger
Cool. Am i internet famous yet? ๐
slender iron
@tulip sleet here is how dictionaries grow in size: https://github.com/adafruit/circuitpython/blob/master/py/map.c#L47
tulip sleet
@slender iron ok perhaps not as bad as I thought. It seems like the first import of any non-builtin .mpy uses up about 3k bytes of overhead, perhaps setting up initial dicts. After that it grows more slowly and in proportion to the .mpy file size. I think I was thrown off by a 4k buffer alloc sometimes happening due to the filesystem, which made it seem like each import was at least 3k. (I had semi-duplicated the file and imported it twice under different names).
slender iron
ya. I'm going to get my gaming rig setup to do more heap captures and render them
bronze geyser
I was wondering how hard would it be for me to add to an m0 Circuit Python: 1) go to sleep. Wake up when motion detected 2) go to sleep. Wake up at 4AM. to circuit python. These are the two features stopping me from using CircuitPython for my sensor projects. Any advice?
prime flower
@slender iron you got a quick plug for the Seattle 3h meetup on Amp Hour. Episode 397.2
bronze geyser
@prime flower I'm enjoying Chris's short interview of folks. Makes me feel like I'm camping along with them.
slender iron
@prime flower yay! I'm hoping I get the experience by meeting folks there
prime flower
ha, absolutely agree, but replace "camping" with sailing as he's doing the latest ones on a sloop
slender iron
@bronze geyser we don't support sleep yet ๐
prime flower
hrm I should start a NYC one, but we already have NYCResistor
there's BarSec but that's not hardware, that's infosec
bronze geyser
@tennewt could i figure out how to write a c module that gets into standby? Then i need to be able to attach an interrupt.?
@slender iron i was looking at the micropython "adding a module" and wondering if i could do something like this? http://micropython-dev-docs.readthedocs.io/en/latest/adding-module.html
tulip sleet
@bronze geyser one possibility for the timed wakeup is to use one of these modules: https://www.adafruit.com/product/3435 or https://www.adafruit.com/product/3573
which could be used to completely power off and power on the board running CircuitPython
slender iron
@bronze geyser yeah adding a module makes sense. I want time.sleep to be smart too but it would be time based
bronze geyser
@tulip sleet Thank you. i will look into that.
tulip sleet
The enable-pin one can control the EN pin on a Feather, etc., which is basically an on/off switch that disables/enables the regulator.
bronze geyser
@slender iron so i wanna see how hard it is to write a module for circuit python. i realize i should just jump in...but in case you have any "getting started" advice?
slender iron
@tulip sleet has a build tutorial and this describes some code layout: https://circuitpython.readthedocs.io/en/latest/docs/common_hal.html
bronze geyser
@slender iron @tulip sleet Thank you.
fading solstice
@raven canopy i still don't see a document in RTD for the l3gd20 library. Do i need to ask Scott to do somehting?
ruby lake
@tulip sleet thanks, I'll have a look
marble hornet
hey, so I just got back and soldered some headers onto the board i made. and it doesn't show at all.
i tried reflashing the uf2 bootloader but am getting a 'can't halt cpu error'
stuck elbow
did you measure the voltages?
marble hornet
yes
it all checks out
i even replaced the reset resistor!
which was shorting
i am trying h, r, r0, rx etc
i have power cycled multiple times
manic glacierBOT
stuck elbow
probably bad solder somewhere
marble hornet
did 3.0 get released? 3.x?
stuck elbow
I think we are at candidate 1
marble hornet
i have local brach that i want to update the changes from it: git fetch https://github.com/adafruit/circuitpython.git?
slender iron
@fading solstice I'll set it up now
raven canopy
@marble hornet the "cannot halt" happens to me occasionally. usually takes a few unplug cycles, then it starts working again. gremlins are ornery...
marble hornet
hey, i think there is a problem with the current master. but it might just be me: i tried:Jonahs-MacBook-Pro:atmel-samd jonahy-m$ make BOARD=feather_m4_express Use make V=1, make V=2 or set BUILD_VERBOSE similarly in your environment to increase build verbosity. make: *** No rule to make target `peripherals/samd/clocks.c', needed by `build-feather_m4_express/genhdr/qstr.i.last'. Stop..
probably me mis-using git
raven canopy
you need to do git submodule update --init --recursive after cloning or updating.
marble hornet
and that does?
stuck elbow
it updates the subrepos
raven canopy
the samd peripherals were moved out so they can be used by makecode in the future
marble hornet
okay
raven canopy
if you look on github, you'll see that folders like peripherals and asf4 are links to separate repos, within the repo. these are known as submodules in git.
marble hornet
i noticed it takes longer to compile now, just size?
raven canopy
i haven't noticed any real increase in compile time. were you using make clean each time before? compiling on an existing build folder (without 'clean') skips the QSTR creation part of the compile.
raven canopy
yeah, the first build would be the same as clean. any builds after that without clean will go faster. but building fresh is highly encouraged when changing the firmware.
@slender iron @idle owl should we do another release for @fading solstice's L3GD20 library since the badge fix and RTD addition?
idle owl
@raven canopy Releases really only matter with code changes, things that would affect the library usage in the bundle. We can probably wait on making another release, and then include those changes in the next release following a code change.
raven canopy
๐ i was thinking more along the lines of asking adabot to use the most recent tag in the bundle and trigger the RTD build. The RTD project is there, but still hasn't been built (returns "does not exist").
would updating the tag on the last release work?
idle owl
That I'm unsure of.
wraith tiger
Well, I've completed my uflash hacking (apart from any further changes ntoll might request).
fading solstice
@idle owl @raven canopy Adabot did run last night after my changes were merged. I think we need to make a release to see the RTD generated and available
manic glacierBOT
oh yeah! what i meant was to make the usbboot keep both up to date if possible. i was following the instructions and using a DK :)
ruby atlas
I did some poking around with dotstars last night. it shouldn't be too hard to add dotstar features to pixelbuf without significant changes.
slender iron
#leeks https://youtu.be/sZo-xscmeOA
raven canopy
@slender iron nice! is the artifacting code or recording related?
raven canopy
this is a no-judgement zone. the code will get there... ๐
slender iron
Yup! It wasn't even playable until last night
turns out the game is much easier slowed down ๐
raven canopy
like the slo-mo bomb on smash bros! ๐
idle owl
@slender iron That's great!
marble hornet
i found a mistake in the amg88xx code and have cloned and corrected the issue, but i cant google well enough to figure out how to set up my local as a sperate branch and use the branch to pull. could someone give me some help please
raven canopy
did you fork the repo?
raven canopy
start by doing that, then clone that repo locally.
raven canopy
Step 1: In the adafruit clone, type git remote add tg_techie https://github.com/tg-techie/Adafruit_CircuitPython_AMG88XX.git (replace tg-techie references as necessary/desired).
marble hornet
remote add?
raven canopy
yep. it will make sense in a few. did you commit your changes yet?
marble hornet
yes
raven canopy
ok
marble hornet
done
raven canopy
Step 2: in the adafruit clone, type git push <tg_techie> <branch_name>. that will push your local changes to your fork
marble hornet
so i also need to clone the adafruit one?
raven canopy
i thought you did?
marble hornet
i cloned mine
raven canopy
"i found a mistake in the amg88xx code and have cloned and corrected the issue"
where are the changes you made?
in a local copy of adafruit_circuitpython_amg88xx (git clone http://github.com/adafruit/), or in a local copy of your fork (git clone http://github.com/tg-techie)?
raven canopy
ok. so you already had forked it.
marble hornet
ah, so now i have a fork it pull requests it automatically! thanks
raven canopy
๐
git/github exist in a magical, mysterious land. at least, it usually seems like it. ๐
idle owl
Not for long! keeps working furiously on the guide
marble hornet
yah its like what chips and electronics were to me two years ago! ** preemptive hug report **
umbral dagger
Like 3d design & printing were to me 6 months ago.
ruby atlas
closer to 9 months ago for the 3d printing for me ๐
git is something I migrated to like ... uh... (checks resume) 11-12 years ago
oh wow, so before github
dotstar support not quite right yet
neopixels still happy ๐
plucky flint
@slender iron wow... that's wonderful to see... am I right in assuming there's something akin to the PyGameZero API going to be available..? #leekandpotatosoup ๐
marble hornet
hey, my m4 based board stopped working . it can boot into metrom4boot but circuitpy doesn't show up
so i am using a different storage chip but it was already supported by circuit python. and it just stopped after being unplugged.
are there any cases or known things like this?
all the voltages check out
stuck elbow
tried with an earlier version?
umbral dagger
@ruby atlas I'm not sure when I started using git. Probably in the early days of rSpec, since we started by using Mercurial for it before switching to git. That would have been ~13 years ago I guess. My github user id is 2020. I think that's from before they opened to the public.
timber mango
.oO(The doctor tells me I have been using git every day, often for hours at a time, but I have no memory of it.)
tgtech try out gitk
marble hornet
how can you clone a previous commit or branch
timber mango
each commit has an sha256sum (or similar checksum).
They can be identified by citing enough of it to make it a unique citation.
marble hornet
what is this magic?
stuck elbow
they are also tagged
timber mango
what operating system are you using?
stuck elbow
so you can clone an older release tag
just type git checkout and press tab to see what you have there...
@timber mango isn't that a personal question? ;-)
timber mango
haha @stuck elbow indeed.
worse, it's a profiling question, as asked.
Still,it is a mistake not to tell the blind barber how tall you are.
marble hornet
๐
timber mango
No offence meant to the blind, the tall, the barber, and users of the words 'still' 'mistake' and especially, 'the'.
solar whale
or the word offense ๐
timber mango
(it's an Irish fence)
solar whale
gitk is nice -- a bit overwhelming
timber mango
The keybindings are (ah..) key .. to its enjoyment. It's very linuxy.
vi keybindings j and k are useful. u and d are as well.
What I like about github is that it does syntax coloring inside the context of a diff, whereas that's beyond the scope of what gitk can render (it has to work with the tcl/tk libraries).
marble hornet
okay so i have an old version download locally but am getting an error! ahh. let me do some googling.
timber mango
$ head -4 /usr/bin/gitk
#!/bin/sh
# Tcl ignores the next line -*- tcl -*- \
exec wish "$0" -- "$@"
$ wc -l /usr/bin/gitk
12385 /usr/bin/gitk
@marble hornet depending on how far back you went, you may have to restore the submodules - git submodule update --init --recursive I do this as a matter of habit every time I clone.
timber mango
$ cat -n /usr/bin/gitk |egrep 18|egrep bind|sed 3q|egrep -v ous
2560 bindkey d "$ctext yview scroll 18 units"
2561 bindkey u "$ctext yview scroll -18 units"
Shouldn't be difficult to customize gitk quite a bit.
solar whale
or break it ๐
timber mango
Hehe. It doesn't have write privileges, afaik. I hope not
This is pretty amazing that someone wrote an app of this size in tcl.
marble hornet
it takes me minutes to compile, hbu?
timber mango
about 138 seconds for circuit python
solar whale
depends on the BOARD
timber mango
Fri 15 Jun 22:29:10 UTC 2018 Fri 15 Jun 22:26:52 UTC 2018
end and begin timestamps during a compile.
solar whale
but I just built Metro M4 express -- took about 138 sec ๐
marble hornet
i'm like 4+ minutes into a pocket now video and it still isn't done (i started video after starting compile)
it didn't work
solar whale
so the compile succeeded - can you get to REPL
solar whale
well - I hate to suggest it, but sometimes USB gets confused on my host system after failed connections - Have you tried rebooting host?
raven canopy
dmesg is available on Mac, yes? since i don't use it, i always forget the *nix commands that are still around...
solar whale
yes -- you need to do sudo dmesg
solar whale
but it give a lot of info! hard to find what you want
raven canopy
can't you filter it? or is that a grep practice? (i rarely used grep when i spent more time in linux)
stuck elbow
in practice you only look at the last few messages anyways
I bet in the end it turns out to be something like a broken cable ;-)
solar whale
it just taht macOS seems to write a lot more to it that linux
marble hornet
i though linux wasn't unix based
stuck elbow
it is
raven canopy
Minix, IIRC.
stuck elbow
no, minix is what you have in the intel management module on your cpu ;-)
raven canopy
oh yeah. so many "based on" lore, these days. ๐
stuck elbow
what we usually call linux is the linux kernel plus gnu tools, and "gnu" stands for "gnu is not unix"
raven canopy
for the record, I run an AMD...
stuck elbow
but the whole thing is very much based on the unix philosophy
at least it was, until systemd came
marble hornet
systemd ?
stuck elbow
the service that runs all other services on linux
raven canopy
it was not a well received addition, from what i've read.
stuck elbow
it's an ugly monolithic blob
very much against the unix philosophy of doing one thing and doing it well
raven canopy
definitely more akin to windows (service host)...
stuck elbow
and ios
solar whale
now you are just being mean ๐
raven canopy
lol. it wasn't meant that way. i don't partake in the L vs W vs M game. I can respect each for what they do and achieve.
solar whale
I just added a chromebook to my collection -- so many ways to get confused ....
raven canopy
@marble hornet did dmesg give you any hints as to why it won't mount?
i thought about a Chromebook, but couldn't justify any use case.
that price point is nice though.
wraith tiger
@solar whale An upcoming (no word on when) update to Chrome OS will add Wayland and a virtual machine running Debian Stretch so that linux apps will run on a Chromebook. It may not be compatible with all Chromebooks, though.
solar whale
@wraith tiger nice! I am having fun exploring what is /is not supported -- too many apps - so little time ๐
wraith tiger
Linux is a Unix "work-alike" based on the concepts of Unix, but not the code.
I've been playing a bit with an HP chromebook. It's nice that android apps are now supported, but it will beceom much more usefull when there is linux app support.
marble hornet
if the error happened after i unplugged it could the bootloader have been corrupted in the section that executes the uf2 but not the one that updates the firmware?
wraith tiger
It is possible to run linux natively on a chromebook, either by replacing Chrome OS or by adding dual boot.
Some chromebooks have upgradable storage, but they seem somewhat rare and are pricier.
Linus Torvalds was a Minix hacker before he started Linux, so that's where the Minix connection comes in.
marble hornet
it magically started working!!!!!!!!!!!!!!! WHAT!
solar whale
ruh- roh
marble hornet
ha-ha
all i did was plug in the jlink
What trickery is this? bit and byte peek-a-bOO?
raven canopy
have you burned the bootloader onto it? (can't remember if we covered that)
marble hornet
we did, with the help of @solar whale
raven canopy
k. so much for "easy answers". ๐
marble hornet
๐คฆ
sharp rain
So I made a python sprite class for animating neopixel strips from BMP files, someone told me Roy might be interested. https://forums.adafruit.com/viewtopic.php?f=60&t=137332&p=679948#p679948
If it's useful I can share to GitHub, but alas... as this is my first python code and haven't done anything with adafruit contrib I would need help on the devops part. Ping me if interested
sharp rain
"Sounds like a cool idea! I'd suggest hopping into the Discord and the #circuitpython-dev channel. (https://adafru.it/discord) Roy there is doing NeoPixel optimizations and would likely want to collaborate on animation libraries."
marble hornet
do j-links do some kinda SUPER hard reset?
raven canopy
@sharp rain Roy is working on the C side with NeoPixels/DotStars. For the Python side, it would be more analogous to a library. And I imagine the community would be quite interested. You can check out this learn guide for how to share a library to the community bundle: https://learn.adafruit.com/creating-and-sharing-a-circuitpython-library/overview
@marble hornet that I'm not sure of. there may be some sort of SWD command it is sending that gets it across some hurdle... ๐คท
raven canopy
so i think i figured out why the M4 is "less" accurate than the M0 for TC frequency capture. The EIC latency. For the M0, it is 3 clock periods (no filter, async). For the M4, it is 5 clock periods. Now...to figure out the math for correction (approxmation).
in my early math, i lose resolution in the upper ranges. ๐ฆ
#stayinschoolkids
manic glacierBOT
marble hornet
i'm looking at the m4 metro schematic and do any of you know what the tcc and other text next to the headers are?
stuck elbow
tcc is the timer controller
it's a bit specific to the samd architecture, it lets you assign timers to pins
stuck elbow
the datasheet says that pcc is the parallel capture controller, so logically, it must be a peripheral that captures signals on a whole range of pins in parallel
it says it can do up to 14 bits at once, which would match the D0-D13 range
marble hornet
thanks!
timber mango
I just did a brief M0 timer with interrupts project today.
https://github.com/wa1tnr/m0_timer_standalone
That'll demo on any SAMD21 target Adafruit stocks, most likely.
(including crickit, haha)
/* ideas sourced from MartinL on forum.arduino.cc */
/* [ http://forum.arduino.cc/index.php?topic=332275.17 ] */
what would be the best way to create enums in CP?
if I do
class MyEnum:
Val1 = const(1)
x = MyEnum.Val1
x is just equal 1 where I would like it to equal/print MyEnum.Val1
indigo wedge
Hmm, sounds icky ๐
stuck elbow
why
indigo wedge
Dunno, just using strings for constants where a int would suffice
stuck elbow
it's converted to an int internally
indigo wedge
hmm, maybe I'm misunderstanding then, could you give me a pseudocode example?
indigo wedge
hmm, then no protection for misspelling when i try to compare?
stuck elbow
protection?
indigo wedge
>>> x == 1
True
>>> x == MyEnum.Val1
True
>>> x == MyEnum.Vat1
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
AttributeError: type object 'MyEnum' has no attribute 'Vat1'
Python is dynamically typed, you get no compile-time type checking
how is that "protection"?
your program still crashes
indigo wedge
yeah, if i used strings then x == "MyEnum.Vat1" would just return false causing a silent bug
stuck elbow
write tests then
or switch to a statically typed language
I've heard Rust is getting SAMD21 support
indigo wedge
heh, that's not that welcoming "oh, you don't quite like one feature, just switch languages" ๐
stuck elbow
it's not like it's just one feature, it's the whole philosophy of dynamic languages
indigo wedge
Never used Rust, heard it's nice though
stuck elbow
it is
indigo wedge
I guess I was just a bit confused cause it seems doable from native code, for example:
>>> y = digitalio.Direction.OUTPUT
>>> y
digitalio.Direction.OUTPUT
So I was wondering if there was a way to do the same from pure python, but maybe not
stuck elbow
sure, you can define a five zillion classes
and define __repr__ for them
but it's wasteful, unnecessary and unpythonic
there was a PEP in recent versions of Python 3 that introduced an enum type
indigo wedge
Hmm, I guess you're right, implementing __str__ would allow me to achieve the print effect that I want but there's little gain from that
stuck elbow
I guess what you want is the ide autocomplete?
indigo wedge
Nah, I get that already with just a class with a const, this was more just to have nice prints while developing and debugging
marble hornet
i am trying to make a pogo based swd programmer and want to make the pads / holes as small as possible. the adafruit website says the the shaft of the pogo pin is 1mm in diameter. is there a rule of thumb on what diameter the drill size should be to insert the pogo pin?
https://www.adafruit.com/product/2429
???
stuck elbow
for bolts I usually leave 0.1mm on each side
you can also do pads and solder the pogo pins horizontally to them
helps to have some sort of rig to hold them straight while soldering then
marble hornet
thanks!
marble hornet
if i was selling a board w/ the uf2 bootloader would i be allowed to use the term "express"? has adafruit copyrighted it?
stuck elbow
as long as you don't use the name "adafruit", you can do anything, afair
rigid path
I have a problem in my circuit python code, I think a second set of eyes could spot the problem without much trouble. Should I post here of go to the Adafruit forum? I have a series of nested if statements that I believe are causing the issue.
solar whale
@rigid path itโs fine to post it here. I wonโt be able to look at it for an hour or so. Others may be available. Sundayโs can be quiet.
rigid path
The code is, I think fairly well commented... In a nutshell, I have an M0 Express Feather, with an OLED and Ultimate GPS Feather Wings all loaded on a Feather Tripler. I use the OLED as a status indicator for the GPS unit, and as individual parts, everything works fine. My final goal in to have the OLED flash a couple of different messages as the GPS is getting a fix. Once a signal is acquired, change the display to a Connected message. Now if the user presses one of the 3 buttons on the OLED, they get a different set of GPS data. Maybe Lat/Long on button A, altitude/speed on button B and other misc data on button c.
My problem is I have several nested if statements and I think everything is going pear shaped around line 84, where I change the flashing message to let the user know a signal has been acquired. Though the GPS pickups a connection, the display stays with one of the connecting messages, and the buttons are not changing anything.
I know the GPS is connected as the LED on the unit slows from 1 flash per second to 1 every 15 seconds, or so.
solar whale
first of all - get rid of the "continue" statement in Lin 81 -- -I htink it causes the code to skip the rest of the code.
unless that is what you want it to do. I may have misread it.
@rigid path gps.has.fix it true, then you want it to do the stuff bleow line 81, correct. I think you wan to use "else:" instead of "continue" and indent! I waiting for my GPS to get a fix to test it.
rigid path
Yes, when it gets a fix, I want to do the rest of the code.
solar whale
@rigid path ok -- simple change -- her is a new version -- note yo have ot hold the button for a second in order for it to be detedcd.
change continue to else
@rigid path seems to work -- I'm sure you will have a few changes to make but it is working
rigid path
You mean the A, B and C buttons?
solar whale
yes - it only checks them once per second.
stuck elbow
use the gamepad module
rigid path
Got it. Wow, simple fix. Thanks.
solar whale
You're welcome - looks good --- I have to run!
@rigid path as @stuck elbow pointed out, there are many ways to handle buttons, If you need very quick response, there is a "gamepad" module the may help. They may not be necessary for your application. you just may want to think more about how you are testing for the buttons. You could test the buttons more quickly than you check for the gps.update. use 2 different counters.... experiment.
stuck elbow
it allows you to read the buttons less often, beacuse it buffers them
solar whale
looks great - I have not tried using it yet, but will now!
will it fit in memory on an M0 with OLED and GPS ?
stuck elbow
it's built-in
so no extra memory
(not sure if it's enabled in non-express boards, though)
solar whale
He is using an M0 express - should be fine.
@stuck elbow Wow - it is very nice - and easy to use. I guess I was not paying attention when it got added. I was thinking it was just for complex gaming stuff. I'll never read a button again!
stuck elbow
thanks :)
marble hornet
is there a way to 100% force halt a atsamd51j cpu?
indigo wedge
Hmm, so when using SPI.configure the speed is called baudrate but to get the speed of the bus one has to use SPI.frequency, seems like an inconsistency in the API, was this done on purpose?
marble hornet
b/c mine keeps on failing to burn the bootloader b/c the cpu can't be halted
proven laurel
If I want to run 4 led strips with the possibility of adding servos down the line is a Gemma the way to go? Or should I use 2 boards one for led strips and 1 for servos.
tulip sleet
@indigo wedge it is kind of inconsistent, but not sure we can change that now. Note that the input baudrate may not match the actual frequency because of clock division granularity. The input parameter should probably be called frequency too.
ruby atlas
Woohoo, 144 element dotstar strip aceleration ~4x with the test dotstar pixelbuf code.
tulip sleet
@proven laurel Gemma only has three data pins, and they are limited in what they can do. You could use a Feather or an Itsy Bitsy and have a lot more pins.
slender iron
@ruby atlas did you see @sharp rain 's earlier post? I recommended you two connect
@plucky flint I'm in the research phase in terms of API. if you could recommend games to port I could start there. From the first few glances I've had at pygamezero it feels too magical to me (WIDTH and HEIGHT specifically.) I've focusing on fantasy consoles atm. May port Celeste to Gameboy too and see what API comes out.
slender iron
@stuck elbow rust stuff is here: https://github.com/wez/atsamd21-rs
indigo wedge
Woo, thanks Scott ๐
I have more changes already brewing ๐ Yesterday I worked on getting SPIM3 working with nrfx on the nRF52840 because that one is the only one that can clock 32MHz, until I realized that I'm using engineering samples that don't have SPIM3 working, need to wait for new chips ๐
indigo wedge
I'm not that far behind you, Adafruit ๐ https://twitter.com/arturo182/status/1011225347566784512
ruby atlas
@sharp rain Sounds very interesting!
ruby atlas
@sharp rain is your stuff in C (to extend CircuitPython), or a Python module?
manic glacierBOT
yah - i think we will have people using both the pca DK and the feather52840 so having two board definitions is good.
for example, i plan on putting together the feather PCB this week but will still be using the DK for testing shields since it is arduino-shape and has JLink built in.
so having both is good :)
here's the rev B Feather52840 pinout:
slender iron
<@&356864093652516868> meeting in a few minutes!
slender iron
๐
ruby lake
indeed lol
cunning crypt
More accurately, I forgot what Monday meant
stuck elbow
full inbox and nothing works
cunning crypt
Hug report: Everyone, as always. Now that things are moved around, I should be able to peek at actual specifics more often
meager fog
@indigo wedge hey those modules look familiar!
@ruby lake m4 feaths being made this week
timber mango
I'm running an Astatic 1104C desk mic today.
hug report: pt @river quest for weighing in on my comments about content discovery on Sunday.
ruby lake
@meager fog wheeee
wraith tiger
Group hug plus everyone who wrote the documentation, forum answers, etc that I consulted while working on uflash.
raven canopy
@jerryn, @Andon, @deshipu for the support they've engaged in on Discord and GitHub. I may have missed mentions for others; I blame "Monday brain". @arturo182 for the nRF cleanup; excited to see that port getting attention. Group hug beyond that!
wraith tiger
ruby lake
offhand q, is there a roadmap for getting ฮผpython's hardware timer stuff into cp?
meager fog
@ruby lake timers are so different form chip to chip, we sorta punted. do you need like precision timer access?
ruby lake
@meager fog I want one timer I can set to say 1ms ticks and call a function from it. time division mulitplex of voltages is 10000% easier with that ๐
polling the timer is also acceptable
cunning crypt
Status Update: Things moved. Many things, including my PC. Plus side, I have a setup I can use for Show and Tell. I also have some EL Wire in the mail I will be doing fun projects. Pic is new computer desk setup.
wraith tiger
IS that the bridge of the Enterprise?
stuck elbow
you forgot your head
cunning crypt
Featuring Creepy Mannequin Head
wraith tiger
Er, bridge, rather
ruby atlas
@cunning crypt how many pixels?
cunning crypt
1920x1080x6 = 12,441,600
meager fog
@ruby lake oh we dont have any callback support at this time to start with :/
cunning crypt
Currently, the top/right monitor isn't connected, but the part to fix that should be delivered... soon.
ruby lake
@meager fog I have other ways to get it to work, will just an eight-channel SPI dac
ruby lake
~just use
ruby atlas
@cunning crypt very cool. i have about 15MP on 3 displays - and still run out.
stuck elbow
twin mica
@stuck elbow looks awesome, nice work!
cunning crypt
@idle owl You needed a screen-ing process
timber mango
If you change the screen resolution to 1024x768 or smaller fonts get automatically bigger (at least in Linux)
stuck elbow
320x200 ftw
wraith tiger
I'd bet you could script screen grab post processing with GIMP.
timber mango
status report:
Did a lot of experimenting with git and github this week, and found a means to synchronize the branch to master (in terms of number of commits ahead or behind) on stackexchange.
https://stackoverflow.com/questions/34118404/what-i-can-do-to-resolve-1-commit-behind-master
Being able to sync allows me to experiment more freely without worrying so much about consequences.
@tulip sleet 's suggestion to try gitk was apropos.
gitk allows me to gain good visualization of the state of my local git repository, prior to any commit.
Looked at diode clamping protection of port pins (again) this week.
I'm in the process of distinguishing placebo from useful protection.
cunning crypt
@timber mango Can definitely understand you clearly. It does make you sound like you're on a HAM radio, but there's no issues in understanding.
wraith tiger
Gui tools are helpfull, too.
twin mica
โCPX touchtone synth with MakeCode project video + learn guide
โNew CPX CAD model update on github https://github.com/adafruit/Adafruit_CAD_Parts
โReleased SVGs on GitHub for iron-on vinyl apparel https://github.com/adafruit/Adafruit_Learning_System_Guides/tree/master/Circuit_Playground_Papercrafts/Vinyls
timber mango
http://adafru.it/1824 for touch-tone pad. ;)
ruby atlas
Dotstar too
tidal kiln
@twin mica maybe model rocket engine ignitors? (to make fireworks go)
rigid path
Can the function of the reset button on an Ultimate GPS Feather Wing (or any wing) be reassigned? Instead of resetting, I want to have the setup write data off to an SD card when pressed. Playing with an M0 Express Feather and an Ultimate GPS and an Adalogger Feather Wings on a tripler.
idle owl
@ruby atlas Super exciting!
cunning crypt
@rigid path Typically, reset buttons are tied into hardware resets
wraith tiger
Looks fancy!
twin mica
@tidal kiln huh good idea, i think i saw some at hobby lobby
tulip sleet
@rigid path it is hardwired to the reset pin: see https://learn.adafruit.com/adafruit-ultimate-gps-featherwing#downloads
idle owl
๐
wraith tiger
cunning crypt
@rigid path If you want to do some hardware hacking, you could make sure the reset pin is physically non-present, and then run a wire to an unused GPIO pin. Not sure which ones that is on the UGPS Wing, though.
timber mango
glossary: GUI graphical user interface 'gooey'
raven canopy
FrequencyIn Module: I abandoned using COUNT for the time being; MP_OBJ_NEW_SMALL_INT was giving me too much grief, post math. Working on accuracy, as I believe it is largely tied to the EIC latency. Currently catching up on all the math I never did/paid attention to. Hug report to Kahn Academy? ๐
Debating on picking up a nRF52840 Dev board. Who am I kidding; it's not a debate....
Little Leslie amp: Spent some time on this over the weekend. Salvaged a surround satellite speaker, which include a 1" tweeter, 3" mid-driver, and crossover. The Crickit's amp drove them very well. Started a SketchUp for the cabinet, and brainstorming on the rotating bits. Fighting feature creep is tough...
timber mango
roy: excellent audio
ruby atlas
@timber mango $10 usb audio adapter with a $200 gaming headset.
timber mango
That mic is excellent!
rigid path
@cunning crypt @tulip sleet @cunning crypt - Thank you for the info. I was just being curious, I think I will just wire in a button.
cunning crypt
That's honestly probably the easiest way
timber mango
If you stack the GPS feather on the MCU (to expose the antenna to the sky) that reset button can save your bacon when you need to flash the target board when it gets into a state where only the RESET button will wake it up.
NPX:TNG
idle owl
<@&356864093652516868> I'll be finishing the Code of Conduct this week. Please feel free to add any further feedback in the next couple of days. Thank you! https://github.com/adafruit/Adafruit_CircuitPython_Code_of_Conduct/pull/3
cunning crypt
I definitely like the idea of a single RGBLED library.
And "Soon" has arrived. Parts have been delivered, I'm dropping out of voice chat. Enjoy, everyone!
idle owl
Have fun @cunning crypt!
@ruby atlas https://github.com/dastels/circuitPython_dotstar_featherwing/tree/master/dotstar_featherwing
solar whale
@tulip sleet @slender iron did the ESP8266 fix also get merged into master - or just 3.x?
timber mango
branch!
ruby atlas
@idle owl oh cool, that uses adafruit_dotstar directly so it would benefit from pixelbuf automatically.
twin mica
later folks, nice chattin'
tulip sleet
@solar whale 3.x only but I will merge over to master
timber mango
Is this meeting listed for showtimes? ;)
solar whale
@tulip sleet thanks!
raven canopy
Thanks everyone!! Finish Monday strong! Naptime must be earned.
wraith tiger
๐
raven canopy
Back to works for me. Later folks.
errant grail
Had to miss the meeting today. That load of decorative rocks just isn't moving itself out of the driveway fast enough.
cunning crypt
@errant grail You were busy with a rock concert?
errant grail
Hugs to the entire CPy team. Working on a CPy Circuit Playground Express learning guide but found that my verbose code couldn't move to 3.0.0 rc without a complete rewrite.
3.0.0rc took away about 4k of program space compared to 2.3.1
ruby lake
back to part 2 of my midi article
errant grail
My guess is that the CircuitPlayground library for 3.0.0 is somewhat bigger. I do like the sound file playback in 3.0.0 better!
wraith tiger
Rocking out the driveway?
indigo wedge
Hi, sorry I'm late, I'm working on porting the nRF to nrfx drivers and continuing to bring up my project. Also I just got 32MHz SPI working on nRF so that's nice ๐
errant grail
The rocks just don't roll themselves into the beds.
slender iron
@indigo wedge awesome! I'm excited for the nrf stuff
errant grail
Giving a talk this Thursday at our Makerspace. They want to see CircuitPython in action.
wraith tiger
BTW, the documentation link on https://github.com/adafruit/Adafruit_Seesaw points to a nonexistant page (https://adafruit.github.io/Adafruit_Seesaw/classAdafruit__seesaw.html) it should point to https://adafruit.github.io/Adafruit_Seesaw/html/class_adafruit__seesaw.html.
tulip sleet
@errant grail I just gave a talk at the Boston Python goup. By far the best part was displaying REPL or editor side-by-side with an overhead camera showing the board(s).
@wraith tiger thanks I will fix that. I saw that but could not find the right link
wraith tiger
You're welome.
errant grail
@tulip sleet Yup. That's the plan. I have a variety of M0, M0 Express, and Metro M4 Express boards to demonstrate along with a wad of neopixels.
indigo wedge
btw can someone add a BT reaction icon? ๐ https://commons.wikimedia.org/wiki/File:Bluetooth.svg
meager fog
@indigo wedge do you have any toughts on https://github.com/adafruit/Adafruit_nRF52840_Bootloader/issues/4#issuecomment-400063538
modest atlas
Dying for a m4 feather here
ruby lake
I need a couple ๐
cunning crypt
@ruby lake Are you making a pillow?
ruby lake
crow feathers ๐
cunning crypt
....Now we need some sort of device that lets several feathers talk to each other. And it needs to be called a pillow.
ruby lake
module playground v2, more analog I/Os, need the memory space of an m4
my key assigner code successfully ported to cp
usual method of setting keyed outputs is to have a timer event pull outputs from result list and stuff them into hardware (DAC, gate) slots. I will have to use direct outputs for now.
slender iron
<@&356864093652516868> Here is the recording from today's meeting: https://youtu.be/a1TgF5gysXc
indigo wedge
Doing just fullscreen fills, i get 22 FPS on the nRF52840 with 32MHz SPI ๐
stuck elbow
@indigo wedge with ili9341?
indigo wedge
yep, 320x240
I modified the WIP C driver to use a 256 byte buffer cause sending one pixel at a time was way slow
stuck elbow
that's decent, though at 32MHz you should have ~200 FPS
no, wait, divide that by 8
25
so close
does it have DMA on the SPI?
indigo wedge
it does but i'm not using it for now
stuck elbow
you would need two buffers for that, probably, to fill one while the other is being sent
though for just filling the screen with a single color, one pre-filled one would be enough
indigo wedge
IIRC nRF DMA has a way to specify linked buffers for SPI, so can have more than two being juggled
stuck elbow
if you could link it to itself, it could be super-fast :P
indigo wedge
With a larger buffer I get 24FPS
stuck elbow
that's probably as fast as it can go
indigo wedge
yeah, there will always be some overhead
I've spent 2h yesterday fighting with SPI to get 32MHz working until I realized I was using the ancient engineering samples where 32MHz mode was broken ๐
solar whale
Yay! Raspberry Pi and Feather M0 Express are talking via Lora !! ๐ป
idle owl
Nice!!
raven canopy
@tidal kiln (& @idle owl) HT16K33 PR review incoming. I've got a couple "nit picky" things. Nothing serious; documentation.
idle owl
Ooh almost missed a sneaky upstream!
ruby atlas
@timber mango btw just getting back to your comment about audio quality: https://www.massdrop.com/buy/massdrop-x-sennheiser-pc37x-gaming-headset and a Syba SM-CM-UAUD (Cmedia chipset).
The mic is at just the perfect distance, and it has a handy dandy volume knob on the headset.
stuck elbow
gaming...
ruby atlas
I got it because it's basically a sennheiser 59x series with a mic on it.
tulip sleet
@indigo wedge are you doing nrf dev on Mac or Linux (or Windows?) hathach and I both are having trouble using the CDC device presented on the PCA10056 port on Linux and I wonder if you're seeing the same thing (if you're using Linux). See https://github.com/adafruit/Adafruit_nRF52840_Bootloader/issues/7#issuecomment-400126713 and above
tidal kiln
@idle owl @raven canopy thanks for review. made some updates. back to you.
wraith tiger
I rebased my fork to the latest uflash version and made a new release: https://github.com/ScottDWebster/uflash/releases/tag/1.1.3sdw0
raven canopy
@tidal kiln @idle owl is there more stuff coming for the HT16K33 library? didn't want to push a release if there is...
tidal kiln
@raven canopy yep. i'm trying to chip away at the other issues as well. not sure what's best for when to draw the line on pushing a release though. probably best just to ask @slender iron
indigo wedge
@tulip sleet do you mean the cdc from jlink or the bootloader?
I'm using Linux Mint and have no problems with the jlink one but have not tested the bootloader for the nrf port yet.
bronze geyser
@solar whale I was listening to the weekly meet...i thought i heard you mention challenges getting SPI on Rasp Pi to work? (I apologize if I misheard). I've been happily using my Rasp pi port of Tony's RFM69 python library using SPI...wrote stuff up in this wiki in case it is of any use...https://github.com/BitKnitting/should_I_water/wiki/RFM69-on-a-Raspberry-Pi
scarlet fjord
I'm getting this.... issue(?) when checking for errors in Mu - is there a limit to the number of characters i can have in a line??
scarlet fjord
ah, ok ๐
another question then; is... this right?
if cond = false is what im doing, why is it telling me i should be doing what im doing?
am i wrong, or is the warning?
Toggles[] is an array, if im understanding this right...
stuck elbow
"==" and "is"
"is" is an operator
but it's best to write if not Toggles[keyindex]: in this case
the == False is spurious
it's like writing if (((x == True) == True) == True) == True:
instead of just if x:
scarlet fjord
ah, fair ๐ thanks
tulip sleet
@indigo wedge I mean the jjlink CDC: connect to the USB port on the short side of the board after CircuitPython is loaded. With screen, picocom, or minicom, I type and get no input echo: nothing happens. If I press reset I see part but not all of the startup prompts. See https://github.com/adafruit/Adafruit_nRF52840_Bootloader/issues/7#issuecomment-400048531. Not sure if this is fixed by newer SDK or it is a jlink software problem. I see this on Ubuntu 16.04 or 18.04 (Mint should be similar). As hathach mentions at the end, this will become a moot point.
indigo wedge
Hmm, that's odd, I connect to the jlink usb connector and use minicom without any problems.
But as mentioned, I don't use the adafruit bootloader just yet so unsure if that changes things
latent nimbus
Good day, I am new here. Just ordered some devices from Adafruit on last thursday and got it yesterday here in the Netherlands. So today I started a test with cirquitpython, and I think it is great. I made a test with a Metro Express M0, a si7021 sensor and a oled 128x64 display. In my program I first write a header and a footer to the display and then in a while true I refresh the data. So I must clean the area for the data, and did that with some for loops. Is that a good way? or is er a better way to controll the oled display: https://github.com/WimAlsemgeest/cirquitpython_M0_temp_hum_oled
tulip sleet
@indigo wedge that may indeed be the problem, but I don't know enough about what the bootloader is doing to understand why that would matter.
@latent nimbus There's an oled.fill() method. old.fill(1) will turn all the pixels on; old.fill(0) will turn all off.
dreamy orbit
hrm...might need some help and very new to programming....I just got a adafruit feather M0 express with the DC Motor + Stepper FeatherWing to control a nema 17 motor for a filament rewinder project. Thinking of using some switches for power, forward and reverse and potentiometer for speed.....how would I do this in circuitpython or would arduino be better?
or should i ask this under a different topic sub channel?
latent nimbus
@tulip sleet Yes I saw that, but I just wanted to clear a part of the display.
tulip sleet
@scarlet fjord [we're moving over here from #help-with-projects].
scarlet fjord
yep thanks ๐
ive just been working on it while you caught up with everyone else ๐
tulip sleet
@latent nimbus ah i see, sorry. We are going to improve the library, which is fairly weak now: we need to do some work on the underlying framebuf library. But your way is fine for now. Or you could draw the outlines once, and then write the proper number of text spaces inside your headers to clear the text and leave the lines alone
scarlet fjord
idk what i did differently, but when i went and changed all my variables to lower-case, it started working ๐
so im plugging in some more buttons to test it properly
tulip sleet
what editor did you say you're using?
scarlet fjord
im working in Notepad++, but im copy-pasting it all to Mu and saving it to the main.py file there because im difficult ๐
tulip sleet
ah, ok, I was going to mention it's important to use an editor that writes completely each time, but mu does: https://learn.adafruit.com/welcome-to-circuitpython?view=all#1-use-an-editor-that-writes-out-the-file-completely-when-you-save-it
scarlet fjord
oh cool, yeah the tutorial said to use Mu so i am, but im not 'working' in it ๐
also follow up on what you mentioned in #help-with-projects - how do i see results from print("")?
like, where does it print to
Mu has a Repl thing, but it doesnt show anything and idk what it is ๐
tulip sleet
@scarlet fjord what version of mu is it?
scarlet fjord
1.0.0.beta.15-adafruit_b142c3, according to the header bar in windows
tulip sleet
there is a newer version (now available at https://codewith.mu), but I believe that one should show the repl. It is the "Read Eval Print Loop", where you can type Python and see the results, and where print statements go.
if you click the REPL button, a smaller subwindow should open at the bottom. Leave it open and that's where the print statements will show stuff
scarlet fjord
oh ok, maybe i just havent seen any of the prints, but ive had that open this whole time ๐คท
scarlet fjord
oh my god is that why its been not working?
ive been copy-pasting
so ive ctrl-c'd a lot
i see the '>>>'
how do i start main.py up again
stuck elbow
ctrl+d
scarlet fjord
that seems to duplicate the current line...
scarlet fjord
eyyyy ok no i did not
i just downloaded the newest version and when i opened the 'serial' window (which replaced the 'REPL' button in the version i was using before) it prompted me to do ctrl-d and then i saw the error i wasnt seeing before!
there was this thing where, on one if the files somewhere, it showed a 'D13' being imported but not used
it didnt throw an error when checking in Mu, but the REPL showed an error for it about an array out of bounds and now it works ๐
tulip sleet
ok, great!
scarlet fjord
still havent truly tested it with more buttons, i only just now managed to pry some key caps off my old keyboard so im plugging them in now ๐
woohoo! it works ๐
four buttons, four keys registered on the PC!
thanks a bunch for your help ๐
stuck elbow
nice switches, are those gaterons?
scarlet fjord
nah, Cherry MX Red RGB ๐
stuck elbow
ah, so Cherry now also makes transparent switches, nice
scarlet fjord
id like to get like... 1 or 2 of each different kind of switch... Cherry, Gateron, Kailh... and see what ones i like before i finish this, but at the moment i just have these ๐
yeah, that keyboard in the background of that shot is all Cherry MX Red RGB
linear + RGB = ๐
though afaik Corsair owns the rights to all cherry rgb keys, which.... sucks, tbh, but it means were getting a lot more competition from other companies, so maybe cherry will up their game haha
stuck elbow
I absolutely love my Kailh PG1350
scarlet fjord
obligatory code screenshot ๐
stuck elbow
one hint, it's traditional in Python to name variables with lowercase, "like_this", class name with camel case, "LikeThis", and constants with all caps, "LIKE_THIS"
it doesn't matter for your own code, but it does make it easier to read for others
scarlet fjord
noted ๐
i dont come from a programming background so this is all very new to me ๐
stuck elbow
that's basically just nit-picking, so don't worry
scarlet fjord
haha ๐
so variables = snake_case
classes = CamelCase
idk what a class is, but im learning ๐
ruby lake
variableNamesAsLongAsYourForearm
scarlet fjord
question:
on this tutorial - https://learn.adafruit.com/arcade-button-control-box/software
is pin D13 used anywhere? i think thats where i was getting thrown off and i cant tell if it was because i messed with it too much or what ๐
Mu complains that 'pin D13 is imported but unused', so I, being me, added pin D13 to the buttonpins array, which of course would break it
stuck elbow
doesn't look like it is, I guess they just imported all the pins
personally I prefer to do import board and then do board.D13 instead of D13 and so on
scarlet fjord
i suppose thats a bit like the argument import time then time.sleep() vs. from time import sleep then sleep()? ๐
scarlet fjord
i can see that.
me though, i like the idea of only importing what I'll use; so im not importing the entire time library just to do sleep, which seems wasteful. unless Python is smart enough to not import things its not going to use, which... wouldnt seem to be the case, if Mu complains about pin D13 ๐
aw, dynobot doesnt like self-deprecation ๐ข
how am i going to do my classic NZ humor ๐ฎ
what i was trying to say is it would be nice if someone edited that tutorial to not import pin D13, cause that confused me when i did 'check' in Mu and saw an error about it and i fixed it the wrong way
sharp rain
@ruby atlas It's a python library that uses the neopixel_write. I'm seeing 1800 pixel / second draw rate on my feather express M0 and memory use is fairly minimal, basically the bytes for the BMP palette and pixels, so for my purposes no need to implement in C
@ruby atlas here's the source code. https://github.com/aaronaverill/CircuitPython_NeoSprite
@ruby atlas it's pretty much done except for more robust examples and documentation. I've used it both with 32 pixel featherwing neopixel matrix for doing simple animations, and on a 150 pixel 5m strip to do animation and chase sequences.
@ruby atlas with feather express and circuit python it detects any changes on the flash drive and auto-reloads so it's wonderful for prototyping - edit the BMP file to change the animation, save, and it instantly reloads and starts playing. Cool stuff.
ruby atlas
That should automatically benefit from the acceleration of pixelbuf. awesome. I'll give it a try tonight when I get home. Thanks so much for sharing that.
At least I believe so.
Oh, I missed that you use neopixel_write directly. Hmm. That means it doesn't do dotstars yet.
I definitely want to look closer ๐
sharp rain
@ruby atlas it's a bit of a lie... I use the NeoPixel python library, but I bypass all the methods and blast the pixels into the internal buffer, then call the show() method that writes it out
@ruby atlas but basically, it's implemented by a method "fill bytes" that will fill any python bytearray with the RGB data in the format of the four channel options, so it's easily adaptable
stuck elbow
@scarlet fjord you are importing the entire library either way
ruby atlas
@sharp rain That makes sense. I'm eager to look at your code later today to see how we could make it compatible with pixelbuf (or extend pixelbuf) so it's portable across device types. Dotstar requires the start/brightness byte before every pixel (plus start 0x0s and end padding 0xFFs).
sharp rain
@ruby atlas fairly trivial to add - there's already support for a channel parameter which specifies the number of bytes to write for each RGB tuple, and how to map the bitmap RGB to which bytes, as well as detection for a W channel if all colors are the same.
@ruby atlas the code should be straightforward. there isn't really any magic. the animation speed comes from:
a. avoiding all the method calls to set the pixel RGB one by one and setting the buffer bytes directly inside the loop
b. avoiding the brightness adjustment calculation inside the loop by tweaking the bitmap brightness once at the start
@ruby atlas I was doing both of these in my first iteration to get it working. When I added these optimizations I saw a 10x factor speedup.
@ruby atlas Moreover, since it's a static array of RGB pixels, you can do things like compute an estimate of the total current required for an animation by adding up all the pixel values once, which it turns out is quite useful.
@ruby atlas And you can create animations in a graphic program rather than in code, which is great when partnering with creatives who aren't techie ๐
@ruby atlas the downside is you can't easily do fully algorithmic animations, but... you've got access to both the pixel buffer and palette as bytearrays, and you can manipulate those. so for example, load an animation file, then tweak the palette to move thru a color wheel. lots of options
rigid path
Circuitpython on non-Express M0 Feathers... with the reduced storage on the non-Express M0 feathers is there a way to tell which library files I need to load? example: I am running a pair of M0 RFM69HCW Radio Feathers. In the sample code I am playing with, I have imports of board, busio, digitalio and adafruitrfm69 in my code. Is there a way I can determine which libraries I need to load to the Feather? Also is there a way to determine which files in the library folder I need to load?
stuck elbow
the documentation for those libraries should mention what the dependencies are
you can also look at their code and see what they import
rigid path
@stuck elbow - Thank you. Are these on read the docs?
stuck elbow
should be in the readme in the library's repo
rigid path
๐
onyx hinge
@rigid path look at the list of modules in sys.modules after all imports are done. I think that works/exists in cp
slender iron
@indigo wedge did you find the nrf micropython PR?
rigid path
@onyx hinge - TY
indigo wedge
@slender iron i didn't check but they don't have the common-hal modules so unsure if it will help much.
slender iron
yeah true
indigo wedge
Also don't want to get too influenced, then I won't know which ideas are mine and which are copied ;)
indigo wedge
Thanks for the merges, I have more on the way.
I'm on the phone now so can't read the pr properly
But will check it
tulip sleet
@scarlet fjord I am changing the example for the Arcade button box so it just does import board. D13 warning will go away and people will be free to change the example without having to change the import list.
thanks for pointing out this annoyance
idle owl
@slender iron Would you take out the _read on this? Or is read not quite like get/set: python def _read_temp(self, register): temp = self._read_16(register) temp *= 0.02 temp -= 273.15 return temp
idle owl
Ok here's the file (still has linting issues):
slender iron
no pr or anything?
idle owl
Not yet, the repo doesn't exist yet
tulip sleet
@idle owl it's an internal routine, not a property in an official API. I think it's fine. It does an action of reading (like, it calls _read_16()
idle owl
That's what we're working on right now
Ok that's what we thought but wanted to make sure
slender iron
I'd suggest making a repo and having a pr for it
tulip sleet
I thought temp meant temporary, not temperature
slender iron
why not do it publicly?
idle owl
No reason other than it fell into place this way
slender iron
can you make repos?
idle owl
Yep doing it now
slender iron
k cool
idle owl
It should be called Adafruit_CircuitPython_MLX90614 right? Formatting-wise I mean
slender iron
yup
tulip sleet
@slender iron @idle owl adding a new library, should I use a webhook for RTD instead of a service? GitHub says services are deprecated. And for travis?
but a lot of libraries still use services
slender iron
yeah, webhook
idle owl
Yes to webhook.
slender iron
travis has some newer stuff (App I think) that we haven't switched to yet
tulip sleet
so keep using service for travis for now?
what's a good example library with the perfect preferred setup?
slender iron
I don't know what you mean by service
usually its just enabling on the travis side I think
tulip sleet
slender iron
it might do that automatically from travis
tulip sleet
ah, ok, so it showed up when enabled on travis, ok, will check that
on the RTD side, it will give me a webhook to paste?
tulip sleet
should we go back and replaces the RTD services with webhooks, or just wait until they say it's going away, fix it?
like in the library above?
slender iron
I would just wait
tulip sleet
or if it ain't broke, don't fix ti
slender iron
right
slender iron
np, glad to get it setup
idle owl
np
tulip sleet
@idle owl https://github.com/adafruit/Adafruit_Learning_System_Guides/pull/232 now suceeded.
idle owl
Merged.
tulip sleet
Tnx. I actually made the fix through the github web editor, since it was so tiny. Rarely do that.
idle owl
Same, but I've also done it for tiny things. It then messes up my local branch though so if it's anything I have locally, I'll usually still go through the process.
tulip sleet
@idle owl so how do I add the library as an RTD subproject? The only project I have in RTD is CircuitPython. I went to Admin->Subprojects, but there's just an empty dropdown for "Child", no place to add a URL.
should I import it as a regular project first?
idle owl
You import it first. On its own. It'll seem like you're importing it into your stuff but you're not
yeah
tulip sleet
ok, got it, I will proceed and whine if it doesn't work.
idle owl
๐ ok
tulip sleet
@idle owl, ok, now it's a subproject, but it's still listed as a top-level project too. ?
idle owl
Yeah it won't show up anywhere but there as a top level. It shows up correctly as a subproject under CP
I think anyway. Because that's what I've been doing.
tulip sleet
so you have a bunch of random library projects and also the CircuitPython project?
idle owl
Yes.
idle owl
Excellent
marble hornet
Thanks you adafruit
tidal kiln
oops. i broke : display support for the 7 segment. this was previously done with some logic. but thinking of fixing it by now making it a property:
display.colon = True
thoughts?
for ref, it was previously done like this:
display.put(":")
and the only way to turn it off was to fill(0), which cleared the whole thing
manic glacierBOT
tidal kiln
hmmm. lemme check. the one i have here only has one. but there may be variation in shop...
looks like only one. and the alphanumeric ones don't have any.
tidal kiln
well, the 1.2" versions have 2 colons. and only one decimal point.
manic glacierBOT
tidal kiln
and it's not really a decimal. more like an AM/PM indicator.
slender iron
tulip sleet
@slender iron @idle owl the code formatting in RTD varies from library to library. In the CircuitPython project and in the Crickit subproject, it looks like this
tidal kiln
@slender iron thanks. will check into that...
idle owl
I think we updated the theme on some of them and not others. I noticed that too.
slender iron
thanks!
idle owl
My guess was theming. But is the second example from a docstring or is that a .md or .rst or whatnot file?
tulip sleet
The difference is that the first is using just theme.css, and the other is using sphinx_rtd_theme.css. It's different defns for <pre>. The first one is actually problematic: the lines are squashed together.
idle owl
Ok so it is theming.
tulip sleet
i looked at the conf.py's but they are virtually identical. one has napoleon, but I removed that and it didn't make a difference. I will force a rebuild on the second one and see if it changes appearance.
idle owl
Ok