#circuitpython-dev
silver tapir
@slender iron Noted, spliting micropython and python.
silver tapir
Date tree ready.
I haven't pull requested to adafruit's git yet. So if anyone has any comments here,
or pull requests on github here: https://github.com/fede2cr/awesome-micropythons
tulip sleet
@idle owl @raven canopy I'm getting this error in GitHub Actions trying to publish a new library release: HTTPError: 403 Client Error: Invalid or non-existent authentication information. for url: https://upload.pypi.org/legacy/. Did I fail to update something in the workflow file? https://github.com/adafruit/Adafruit_CircuitPython_BLE_Cycling_Speed_and_Cadence/runs/443818647
raven canopy
@tulip sleet did you add the pypi user and pass to the repo's secrets? They should be named TWINE_USERNAME & TWINE_PASSWORD.
tulip sleet
no, but it didn't do that on a previous repo either (heart rate) and it was ok (but maybe someone else did) but in any case, I'll do that, thanks
raven canopy
Hmm. Is the repo supposed to be on PyPI?
tulip sleet
no, not yet
i'll compare the actions with heart_rate
heart rate is on pypi, and ancs is too, but not other ble libs. I'll try to make this all consistent
we are interested in getting ble to work under blinka, so we may as well put them on pypi
raven canopy
Hmm. I wonder if we should add a conditional that gives a better message for missing twine secrets... ๐ค
tulip sleet
@raven canopy I'm looking in the "adding a library" Learn Guide and I don't see anything about twine stuff either in the public pages or limor's private checklist. Also that guide still talks about travis. I removed some travis info from the checklist
all set now
raven canopy
Sorry about the secret name mixup! Need more coffee. ๐
prime flower
@tulip sleet do we have a port of circuitpython for the nRF52832
tulip sleet
not anymore, we had a weak version maybe in 3.x or early 4.x
it's too small
and no native USB
prime flower
:\ dang, got a RAK4600 module in for evaluation (NRF52832+SX1276 LoRa module) and was wondering
thanks!
tulip sleet
:wish it were better news:
prime flower
np - would've tried porting it this weekend. wanted to check first
tulip sleet
@prime floweru I was just updating the bundle doc, we have these overlapping sections:
Internet of Things Web Service Helpers
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Helpers for connecting with hosted and self-hosted internet-of-things web services.
.. toctree::
Adafruit IO <https://circuitpython.readthedocs.io/projects/adafruitio/en/latest/>
Amazon AWS IoT <https://circuitpython.readthedocs.io/projects/aws_iot/en/latest/>
Azure IoT <https://circuitpython.readthedocs.io/projects/azureiot/en/latest/>
Google Cloud IoT Core <https://circuitpython.readthedocs.io/projects/gc_iot_core/en/latest/>
LIFX Lights <https://circuitpython.readthedocs.io/projects/lifx/en/latest/>
Hue Lights <https://circuitpython.readthedocs.io/projects/hue/en/latest/>
Internet-of-Things/Wireless/Bluetooth/BLE Protocol Helpers
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Helpers for interfacing with the internet or wireless communication.
.. toctree::
Bluefruit LE Connect App <https://circuitpython.readthedocs.io/projects/bluefruitconnect/en/latest/>
Bluetooth Low Energy (BLE) <https://circuitpython.readthedocs.io/projects/ble/en/latest/>
TinyLoRa TTN Helper <https://circuitpython.readthedocs.io/projects/tinylora/en/latest/>
Mini ESP Tool <https://circuitpython.readthedocs.io/projects/miniesptool/en/latest/>
MiniMQTT <https://circuitpython.readthedocs.io/projects/minimqtt/en/latest/>
JSON Web Token (JWT) <https://circuitpython.readthedocs.io/projects/jwt/en/latest/>
Requests <https://circuitpython.readthedocs.io/projects/requests/en/latest/>
NTP (Network time Protocol) <https://circuitpython.readthedocs.io/projects/ntp/en/latest/>
should we merge this or move some stuff from the second IoT into the first IoT and change the title of the second? I was thinking of splitting off a BLE section but you were talking about interoperating BLE and MQTT support later
i guess "service" and "protocol" are slightly different, but not in the BLE world
prime flower
@tulip sleet Service and protocol are different in the IoT-world. Possibly "Services" is a better term?
tulip sleet
well, they are different in BLE really, but we have "BLE Services". This list of BLE stuff is only going to grow: it probably deserves its own section
the title of the second section is misleading: it's actually just "Internet-of-Things/Wireless Protocol Helpers" right now. I was in the middle of editing.
i'm not sure I want to lump BLE and Wireless together, and NTP, etc. are not only "Wireless:
tulip sleet
is LoRa also "wireless", lumped in with http, mqtt, ntp, etc?
it's "wireless communication"
prime flower
We might want to do something like: a section for each transport type (BLE, LoRa, LoRaWAN, Ethernet, WiFi). Then a section for our protocol implementations - like NTP and MQTT, since those will be agnostic to transport
tulip sleet
maybe let's wait to split that later, because right now there's only the one lora helper. but I have more than four for BLE. We can rethink this after the next round of IoT and LoRa libraries
also helpers vs drivers : there's a "Radio" drivers section
prime flower
Yeah, saw that. tinylora is a wrapper over a driver (rfm9x), i'd consider that a helper.
tulip sleet
ok, i will do something or other
manic glacierBOT
ionic elk
thanks @half sedge, PWMOut is fully supported on STM32 already but I couldn't find a nice control program for PulseOut, so that double Circuit Playground one is super useful.
slender iron
@tulip sleet how did you test ibeacon and eddystone? was it useful? it looks like most of the auto-url stuff was discontinued
ionic elk
@slender iron so, on atmel-samd, are timers not actually mapped to pins? It seems all the existing ports just shuffle through a list of timers and pick one that's unallocated for the PulseOut interrupt counter system.
slender iron
a single pin can have multiple timers so there is a list to go through
ionic elk
I don't see anything that checks the "carrier" pwm object before the timer is selected though
So I'm having trouble understanding what the pool of timers is
slender iron
ah, pulseout's pin association is already done through the given pwmout
the second timer for pulseout is only used for the interrupt
pulseout just uses the pinmux to turn off and on the pwm carrier
ionic elk
right, it's that second timer, the one that's found in the // Find a spare timer. section
what pool of timers is that picked from? If I just pick from a list of timers on STM, I'm going to lock big swathes of pins out of having PWM
slender iron
we look at all of them but start with the ones that impact the fewest pins
ionic elk
so tc_insts[index] is some kind of built-in array of timers that's ordered in terms of their pin impact?
slender iron
thats what you could do. samd has tccs for complex timers so any tc is ok
tulip sleet
@slender iron I used some iPhone apps for testing; I still have "nRF Beacons" and "Locate" on my phone so those were probably the most useful. Many beacon apps have heavy filtering, so you have to use the Apple company id and/or give specific UUID's in advance in order to see the beacons. Given that the whole beaconing trend seems to have become moribund, we could revisit whether to fix the implementation of these or drop it for now.
slender iron
ya, @meager fog and I talked about fixing it up for a quick demo by @split ocean next week. I'm thinking getting the radio library released would be more useful
ionic elk
@slender iron I could use timers 6 or 7, which have no pins and are restricted to internal use or DAC use - but is there a chance we'd need those for audio?
slender iron
just use them for now and we can rework it later if needed
timber mango
@gilded cradle PR #60 is the correct one. I closed PR #59.
timber mango
@trim elm My PR on Adafruit_Circuitpython_MSA301 seems to be in conflict with one that was opened before it. Did I do something wrong?
onyx hinge
{ MP_ROM_QSTR(MP_QSTR_dot), (mp_obj_t)&linalg_dot_obj },
``` I think I need to figure ut the difference between `MP_ROM_QSTR` and `MP_OBJ_NEW_QSTR`#define MP_ROM_QSTR(q) {MP_OBJ_NEW_QSTR(q)}
hm
trim elm
@timber mango I closed it because someone else had already opened a pr doing the same thing but they needed to also update examples.rst. Doesnโt seem like theyโre going to do that so Iโll merge yours
timber mango
@trim elm Thank you. I did both parts.
orchid basinBOT
timber mango
When a PR is merged, does the person who did it get an email?
lapis hemlock
tulip sleet
@timber mango if they are set up to receive notifications, yes. See https://github.com/settings/notifications. Also you can "watch" particular repos: https://github.com/watching. See the "Watch" button to the right of the repo name at the top
slender iron
@lapis hemlock nice! I'm happy to look at the changes whenever they are ready
lapis hemlock
@slender iron will probably be tomorrow.
I wanted to point out that here were are going to lose compatibility with numpy...
slender iron
๐ I'm gone all weekend (we have a holiday monday)
would it still work if the imports were import numpy as ulab import numpy as fft import numpy as poly ?
lapis hemlock
That's actually right. I have never seen such an import.
slender iron
๐ I thought so
it is definitely weird but it does give us compatibility if we want it
lapis hemlock
@slender iron We have to reckon that the MCU is smaller than a computer. I will just wrap up the code along these lines.
How do you insert code blocks into a discord message?
ionic elk
@slender iron I found an ST appnote that I think outlines the optimal way to implement PulseOut: see "Arbitrary waveform generation using timer DMAburst feature". Could you give it a quick scan and see if it's reasonable? https://www.st.com/content/ccc/resource/technical/document/application_note/group0/91/01/84/3f/7c/67/41/3f/DM00236305/files/DM00236305.pdf/jcr:content/translations/en.DM00236305.pdf
slender iron
I'm not sure how it relates to PulseOut. Can't you just copy how samd does it?
nrf does it the same way as well
ionic elk
I'm just confused by what you're doing - you have a PWMOut, but it's just used to pull a pin high and low? And there's a second, arbitrary timer, that isn't mapped to a pin, that's used to time out the intervals?
Why not just use one timer and a digitalIO?
onyx hinge
Use Ctrl-D to exit, Ctrl-E for paste mode
>>> import ulab
>>> import ulab.matrix
>>> ulab.matrix
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
AttributeError: 'module' object has no attribute 'matrix'
>>> import ulab.matrix as ulab_matrix
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
AttributeError: 'module' object has no attribute 'matrix'
>>> from ulab.matrix import inv
>>>
``` hmm something about builtins with submodules doesn't quite work as expected.@lapis hemlock you surround the code in triple backquotes for multiple lines. For a single word or line, use single backquotes.
basically like markdown in github
lapis hemlock
Oh, so exactly as in github. Well, then for the rest of the gang, this would have been my first post today: ```python
import ulab
from ulab import fft
from ulab import poly
a = ulab.array([1, 2, 3, 4])
poly.polyval(ulab.array([1, 2, 3]), fft.spectrum(a))
@ionic elk IR has a 38khz carrier wave that the PWMOut generates
ionic elk
Wait so it's not actually a pulse at all, it's a pulse train on an interval?
slender iron
tulip sleet
you can add a language designator after backquotes like "python" to to get it color-coded
lapis hemlock
@onyx hinge Jeff, you have to do the import as in the earlier post.
ionic elk
Ok, I get it now, I missed the carrier instance, I thought these were literal pulses
tulip sleet
immediately after the triple backquotes
ionic elk
"N-pulse waveform generation using timer synchronization" then
tulip sleet
you can edit the previous post by clicking on the three dots you get when you hover over the post
lapis hemlock
Doesn't seem to work. Well, whatever.
timber mango
@tulip sleet OK, thank you.
lapis hemlock
@tulip sleet there must not be a space after python, that was the problem.
tulip sleet
same is true in github
lapis hemlock
@tulip sleet ๐
slender iron
@ionic elk please just copy how atmel and nrf do it
grim jetty
I'm working through the basic tutorial for adding a C-code module to circuitpython. The directions seemed good until I got to the part about modifying the ports/atmel-samd Makefile. I made (what I think was) a good guess, but then the next step of modifying the mpconfigport.h file has me completely lost. I can't find anything in that file similar to what the instructions suggest will be there. Can anyone guide me through this part? @slender iron? Thank in advance!
slender iron
@grim jetty would you mind posting your modifications to github? I'm going to grab lunch and will get back to you after. @tulip sleet we should update @umbral dagger guide since the CIRCUITPY_ rework.
tulip sleet
I'll take a look and do some editing!
lapis hemlock
@onyx hinge @slender iron I have committed my code to https://github.com/v923z/micropython-ulab/tree/2dim/code. I still have to move three functions to ndarray.c, but I haven't time today. In any case, if you have comments on the code, let me know!
onyx hinge
fast work! I will try to take a look tomorrow.
lapis hemlock
OK, thanks! I will clean up the three remaining functions, and fix the issue that we discussed on github over the weekend.
lapis hemlock
A question, unrelated to anything that I have mentioned so far. There is a wish list here: https://github.com/v923z/micropython-ulab/issues/14 What should one do with requests that have been implemented? Delete from the list, or leave it there for future reference?
grim jetty
@slender iron I'm following guide here: https://learn.adafruit.com/extending-circuitpython/inside-the-virtual-machine but I lose sync between the actual code I'm finding versus the instruction when I get to the point where I'm supposed to edit the Makefile and the mpconfigport.h file. So my code is (hopefully) just the example in the tutorial up until the point where the instructions don't match the v5 source code anymore.
tulip sleet
@lapis hemlock you can consolidate all the ideas in a checklist in the top post. If you put "[ ]" at the beginning of each item, they turn into checkboxes. https://help.github.com/en/github/managing-your-work-on-github/about-task-lists
or you could open an issue for each major addition and close it when added
lapis hemlock
@tulip sleet I like the idea of checkboxes, but the second proposal is also sound. It is still OK, though, the list is getting long and somewhat chaotic, hence my question.
tulip sleet
we tend to open an issue for each suggestion. Then there can be a thread for a suggestion and it can be discussed if necessary. And when a PR fixes it or it gets rejected, we can just close it
I don't think we've done a lot of task list issues. If we use them, it's for subtasks to implement a single major feature.
lapis hemlock
@tulip sleet But if you open a new issue for each item, then the wish list is rendered redundant, isn't it?
tulip sleet
yup, so you could edit each suggestion, point to an issue, and then close the wishlist. In the future each wish can be its own issue
@grim jetty take a look at py/circuitpy_mpconfig.h and py/circuitpy_mpconfig.mk. These consolidate naming features and turning them on and off across ports.
lapis hemlock
@tulip sleet My initial idea was that having a sticky issue at the very top would encourage people to toss ideas. I don't know, whether that has materialised.
tulip sleet
@grim jetty also py/circuitpy_defns.mk
we have an enhancement label we put on suggestions, and we also assign it a milestone, usually 'Long Term', or the upcoming next release.
your sticky wishlist issue could be locked, but with instructions to open a new issue, and with a GitHub search URL to show all the 'wishlist' or 'enhancement' issues.
i think we feel issues are cheap, it's fine to have plenty
lapis hemlock
Sounds good.
slender iron
@lapis hemlock diff looks good to me!
grim jetty
@tulip sleet I think I managed to get a new teensy40 image built with the demo module (but can't test it until I get home tonight), but there were some other issues with the online tutorial in addition to the ones mentioned.
timber mango
I am reading the docs for displayio. There does not seem to be a sample.bmp file provided for the example.
half sedge
@half sedge The library was updated an hour ago with what I believe is a fix to that issue. We haven't released it yet until it's tested more thoroughly, but you can get the updated .py file from the repo.
@idle owl adafruit-circuitpython-bundle-5.x-mpy-20200213.zip does not contain the corrected version, maybe because a new release should be created: https://github.com/adafruit/Adafruit_CircuitPython_Display_Text/releases ... I am safe because I now use label.py on all my board.
half sedge
If anybody has a PyPortal Titano handy, I would like to test a piece of code to see if I have the right scale_factor (I used 12) for that screen resolution. It does not require any other hardware. https://github.com/dglaude/CircuitPython_MLX90640_PyGamer_Plus/blob/master/nomlx_scale_pyportal_titano.py Thanks in advance. (I also need to know your feeling about font readability on that screen...)
river quest
The new Adafruit Bluefruit Playground app is out! Here's a preview of Bluetooth Wireless puppets! https://apps.apple.com/us/app/bluefruit-playground/id1489549571 ... https://youtu.be/EhiJ-OzZh1Y
timber mango
The example on this page does not work. I have a version that works on my PyGamer now.
tulip sleet
@timber mango Could you file an issue about this, with your working version? Thanks! https://github.com/adafruit/circuitpython/issues
timber mango
@tulip sleet Sure!
tulip sleet
Check https://circuitpython.readthedocs.io/en/latest/shared-bindings/displayio/OnDiskBitmap.html just to make sure it wasn't already fixed.
timber mango
Nope, it has not been fixed.
manic glacierBOT
onyx hinge
@slender iron @lapis hemlock I'll work on pulling in the module changes and reorganizing the docs in cpy tomorrow morning (us time)
manic glacierBOT
manic glacierBOT
grim jetty
Quick followup question on incorporating an external C-code module into the circuitpython build. I think I got further yesterday with @tulip sleet 's suggestion that some of the edits required are now moved to py/circuitpy_mpconfig.h instead of the old places still listed in the tutorials. However now I'm getting a new error compiling the "mymodule" demo code. As far as I know this is copied exactly from the tutorial. Any thoughts on where I should go from here:
CC ../../shared-bindings/mymodule/init.c
In file included from ../../shared-bindings/mymodule/init.c:3:
../../shared-bindings/mymodule/init.c:11:19: error: 'MP_QSTR_MyClass' undeclared here (not in a function); did you mean 'MP_QSTR_Flash'?
11 | { MP_ROM_QSTR(MP_QSTR_MyClass), MP_ROM_PTR(&mymodule_myclass_type) },
| ^~~~~~~~~~~~~~~
../../py/obj.h:168:56: note: in definition of macro 'MP_OBJ_NEW_QSTR'
168 | #define MP_OBJ_NEW_QSTR(qst) ((mp_obj_t)((((mp_uint_t)(qst)) << 3) | 0x00000006))
| ^~~
../../shared-bindings/mymodule/init.c:11:7: note: in expansion of macro 'MP_ROM_QSTR'
11 | { MP_ROM_QSTR(MP_QSTR_MyClass), MP_ROM_PTR(&mymodule_myclass_type) },
| ^~~~~~~~~~~
make: *** [../../py/mkrules.mk:55: build-teensy40/shared-bindings/mymodule/init.o] Error 1
(Ok, I see pasting error messages doesn't work exactly right because discord things some of the characters are for formatting, is there a way to post 'literal' stuff here?)
raven canopy
you can use a code block, with 3 backticks before and after:
```
text
```
text
lone axle
is there an escape or something? how did you get it so the first text didn't get put into the code box?
grim jetty
Ok, let me try again to post my error message trying to build my example c-code module from the tutorial:
CC build-teensy40/autogen_usb_descriptor.c
CC ../../shared-bindings/mymodule/__init__.c
In file included from ../../shared-bindings/mymodule/__init__.c:3:
../../shared-bindings/mymodule/__init__.c:11:19: error: 'MP_QSTR_MyClass' undeclared here (not in a function); did you mean 'MP_QSTR_Flash'?
11 | { MP_ROM_QSTR(MP_QSTR_MyClass), MP_ROM_PTR(&mymodule_myclass_type) },
| ^~~~~~~~~~~~~~~
../../py/obj.h:168:56: note: in definition of macro 'MP_OBJ_NEW_QSTR'
168 | #define MP_OBJ_NEW_QSTR(qst) ((mp_obj_t)((((mp_uint_t)(qst)) << 3) | 0x00000006))
| ^~~
../../shared-bindings/mymodule/__init__.c:11:7: note: in expansion of macro 'MP_ROM_QSTR'
11 | { MP_ROM_QSTR(MP_QSTR_MyClass), MP_ROM_PTR(&mymodule_myclass_type) },
| ^~~~~~~~~~~
make: *** [../../py/mkrules.mk:55: build-teensy40/shared-bindings/mymodule/__init__.o] Error 1
@lone axle yep. you can escape characters with the usual \.
fallen anvil
@indigo wedge If you think that's bad, wait till you see what's coming next ๐ So...is there a coding standard or similar around?
meager fog
@indigo wedge if/when ya are around and can send me new binaries to try on my feather - i'd try it out!
fallen anvil
I think memcpy would have the same alignment issue I'm afraid, although I doubt it could be caught by the compiler since it's a void * ... raises an interesting question!
indigo wedge
@meager fog will do in a few, i can first send you the ones I made for 4MB flash and I think they should also work for 8MB, if not, I can modify the build
indigo wedge
@fallen anvil not sure if there's a documented standard, I kinda went by looking at existing code in the samd port
fallen anvil
Ah...memcpy deals with the unaligned copy...see lib/libc/string0.c . I'll tidy to be more inline with what else is in the file(s) and resubmit.
idle owl
@gilded cradle Thanks for testing the PyBadger updates!
gilded cradle
Yw, @lone axle was helpful too.
slender iron
@fallen anvil thanks! we'll have a formatter soon. there is a discussion upstream atm: https://github.com/micropython/micropython/pull/5631
tulip sleet
@slender iron I merged the service data PR> I'll make a new release of the BLE library unless you want to wait.
tulip sleet
done
tulip sleet
is it raw asyncio or is it curio or trio?
slender iron
raw
kinda silly that it's all async but still waits to return the results until a timeout is hit
lapis hemlock
@slender iron @lapis hemlock I'll work on pulling in the module changes and reorganizing the docs in cpy tomorrow morning (us time)
@onyx hinge You mean the code from master, right? I have a problem in the2dimbranch. If I have attributes, I can't attach own methods tondarray: ```python
import ulab
a = ulab.array([[1, 2], [3, 4]])
print(a)
a.transpose()
how does it fail?
lapis hemlock
Here is the traceback: ```Traceback (most recent call last):
File "/dev/shm/micropython.py", line 6, in <module>
AttributeError: 'ndarray' object has no attribute 'transpose'
are your changes somewhere? that is likely an issue with the ndarray type definition
lapis hemlock
Here is my locals_dict: c STATIC const mp_rom_map_elem_t ulab_ndarray_locals_dict_table[] = { { MP_ROM_QSTR(MP_QSTR_flatten), MP_ROM_PTR(&ndarray_flatten_obj) }, { MP_ROM_QSTR(MP_QSTR_reshape), MP_ROM_PTR(&ndarray_reshape_obj) }, { MP_ROM_QSTR(MP_QSTR_transpose), MP_ROM_PTR(&ndarray_transpose_obj) }, // { MP_ROM_QSTR(MP_QSTR_sort), MP_ROM_PTR(&numerical_sort_inplace_obj) }, };
slender iron
that looks right. what does dir(a) show?
lapis hemlock
The method is not shown.
slender iron
but it shows the other two?
lapis hemlock
No. That was just an example. All methods from the local dictionary are gone.
Weirdest thing.
slender iron
setting attr might be the problem
lapis hemlock
I can call module methods all right, it is a problem with the locals_dict.
setting attr might be the problem
@slender iron Exactly. If I uncomment the line here https://github.com/v923z/micropython-ulab/blob/d8bfe46bd8a2d5bdb90521f4bcc8a68325a5f21a/code/ulab.c#L56 then it works again.
I mean, if I comment the line out.
lapis hemlock
Probably that is the way to go.
lapis hemlock
By the way, what is the difference between properties and attributes. Beyond the fact that attributes obviously mess something up.
The first sentence was a question:)
slender iron
properties are the mechanic to have attribute access call functions
they are a special case of data descriptors which let you factor out the functions called on attribute access
functions on an object are really just attributes whose values are bound function objects
lapis hemlock
I do exactly this https://github.com/micropython/micropython/blob/10709846f38f8f6519dee27694ce583926a00cb9/py/objslice.c#L66, and this is why I am a bit baffled.
In any case, I will patch up the code with properties, and hopefully that fixes the issue.
slender iron
the locals dict lookup is done by the default attr implementation
so by overriding the .attr you are ignoring the locals dict
lapis hemlock
Oh, well, that explains something...
slender iron
slice can do that because it only has plain attributes
slender iron
this is the default impl: https://github.com/micropython/micropython/blob/10709846f38f8f6519dee27694ce583926a00cb9/py/objtype.c#L581
it's much better to use it rather than override it
lapis hemlock
Darn it, I completely overlooked this!
slender iron
๐ np. it's complicated
lapis hemlock
I might need the weekend to fix everything. I don't know what your timeframe is.
slender iron
no rush
lapis hemlock
OK. I have to dash off now, but I think we have sorted out the problem. Thanks again!
slender iron
thanks @lapis hemlock !
timber mango
@gilded cradle I should have specified that I ran it under 5.x.
I could have been much more specific in my description of the issue too.
onyx hinge
@lapis hemlock I used properties in my wip after seeing a similar problem
Like this: https://github.com/adafruit/circuitpython/pull/2583/files#diff-21cd947fd3bd1d345da8a0e5a9d0d07cR527
gilded cradle
@timber mango, I figured. I already submitted and merged a PR to fix it.
timber mango
Thanks. How do I link a PR to an issue?
idle owl
@timber mango In a GitHub comment type # and it'll bring up a list, type the number of the PR, and it'll reference it.
slender iron
GitHub recently added a way to do it on the right side as well
idle owl
Click on the Preview tab if you want to verify it took the number you provided.
Ah nice.
slender iron
under "Linked Issues"
timber mango
@idle owl I just tried that on an issue but the PR does not show under Linked PRs.
idle owl
I didn't know about that feature until now. If you look at the PR, it should show the issue as having referenced it.
I don't know how the "Linked" feature works yet.
gilded cradle
@timber mango, just paste a url link in a comment and github handles it.
timber mango
Oh, OK.
idle owl
@slender iron Am I seeing right that gamepadshift isn't included for every board?
specifically CLUE at the moment.
slender iron
I assume so
shift should only be needed if the buttons are connected through a shift register
idle owl
@slender iron Ok thanks. Trying to adapt something to CLUE and wanted to understand what I was working with.
slender iron
did you get it going?
idle owl
yeah, it's ugly but it works. Now I have to figure out the best way to do it.
orchid basinBOT
idle owl
@tulip sleet Do you have a few minutes to discuss some code?
tulip sleet
๐
idle owl
vid?
tulip sleet
sure
lapis hemlock
Like this: https://github.com/adafruit/circuitpython/pull/2583/files#diff-21cd947fd3bd1d345da8a0e5a9d0d07cR527
@onyx hinge OK, I will just merge these into2dim.
@onyx hinge I think the code is relatively straightforward now, but I would like to fix the docs before merging everything into master.
Speaking of docs, do you want to sync it with mine, or would you rather do that separately for adafruit?
I see that you do most of it in-line, correct?
umbral dagger
Anyone have insight on this issue? I've veryifired the described behaviour. It crashes hard, dismounting CIRCUITPY. Seemingly at arbitrary times. I've seen it crash in the middle of the inner nested for loop. I used 5.0b5 and yesterday's bundle.
https://forums.adafruit.com/viewtopic.php?f=60&t=162230
It'll crash even without any interaction on the keypad.
lapis hemlock
@onyx hinge it seems to me that "py/objproperty.h" is not in micropython https://github.com/micropython/micropython/tree/master/py. That must be a circuitpython file.
slender iron
@lapis hemlock we may have had to make it public. I bet the types in it are available somewhere. They don't use properties in their native api because they are slightly slower
lapis hemlock
@slender iron We will have to include/not include, depending on the value of the CIRCUITPY flag. Well, what the heck! Or, perhaps, what the hack!
I could also ask micropython to make that public.
It is only the typedef: https://github.com/adafruit/circuitpython/blob/master/py/objproperty.h
slender iron
it is in a .c file in mp: https://github.com/micropython/micropython/blob/e83ab7374bf1e8ee7b0644eb35567e30e6c21c36/py/objproperty.c#L34
lapis hemlock
Right, but then we can't include it.
The funny thing is, I can't find, when it was removed from micropython. From https://github.com/adafruit/circuitpython/blob/master/py/objproperty.h it appears to have been part of micropython earlier.
I wanted to see what their rationale was.
lapis hemlock
But the comments of https://github.com/adafruit/circuitpython/blob/master/py/objproperty.h are from MP proper, aren't they?
slender iron
I just preserved them
lapis hemlock
And this issue is really not ulab-specific. Any external module might need properties, so I would claim that there is a pressing reason for moving the typedef into a header file.
I just preserved them
@slender iron Oh, I see.
lapis hemlock
@onyx hinge Jeff, does python import ulab a = ulab.array([1, 2, 3]) a.itemsize work? I get <property>, nothing more, and I know that the corresponding function is not called at all, because I have a printf in that, and I don't see anything.
At least, transpose is not missing anymore, but I can't properly attach the properties.
slender iron
what is your attr set to? anything?
are you testing on circuitpython or micropython?
lapis hemlock
.attr is commented out. I am testing on micropython.
slender iron
micropython may have property support commented out... I'm not sure
lapis hemlock
That must be the case. Really odd.
slender iron
totally possible we fixed native properties in circuitpython
lapis hemlock
Can you show me how you brought that in?
The fact that there is no header file in micropython is a smoking gun...
lapis hemlock
I will ask on the forum what they think about it.
slender iron
its probably the pr where I add that header
lapis hemlock
OK, I will check.
timber mango
I am running my boards from a Raspberry Pi 4 which is running from its recommended power supply. Is there a reason they should get into safe mode when I connect them? Usually they are both just fine.
How do I get out of safe mode without changing a power source?
tulip sleet
@umbral dagger I am trying it
onyx hinge
@lapis hemlock yes the properties do work in circuitpython. ```>>> ulab.zeros(3).shape
(1, 3)
if it turns out that we have to make this different in circuitpython, then the code structure where all the "bindings" go in your ulab.c works great for us, we can put it in our shared-bindings/ulab/__init__.c . but, is the payoff for having properties in circuitpython's ulab worth making code not portable between them? I'm less convinced. Would want to know what @slender iron thinks
ivory yew
Has anyone gotten https://github.com/myelin/arduino-cmsis-dap to work with macOS?
tulip sleet
@umbral dagger I answered in forum thread; I'm pretty sure it's excess power consumption
tulip sleet
I added some info to the board Learn Guide FAQ about this.
ivory yew
nvm I got another programmer working
is there an easy way to convert uf2 to elf/hex?
nvm I'm dumb
ignore me
I'm very excited. Sol's production sample came today.
onyx hinge
@ivory yew yay!
golden sail
quick question, I connected an esp8266 to a itsybitsy m4 express
im trying to get data from sensors connected to the itsy bitsy m4 and collect them with the esp8266 to later upload them on an app
how would I go about doing this?
ivory yew
onyx hinge
@golden sail https://learn.adafruit.com/adding-a-wifi-co-processor-to-circuitpython-esp8266-esp32/firmware-files may be of use to you. you would use the "AT command firmware". #help-with-circuitpython is a better place to ask if you get stuck. Good luck with your project!
ivory yew
oh look, you'll be able to buy Sol soon. ๐ https://twitter.com/theavalkyrie/status/1228506521723531264
river quest
here's the DRAFT of the upcoming โPython for microcontrollersโ newsletter that ships out tuesday at 11am ET via adafruitdaily.com - please send any links, news, events and more via an issue/PR, or @ us or really any thing ๐ we'll get it added! https://github.com/adafruit/circuitpython-weekly-newsletter/blob/gh-pages/_drafts/2020-02-18-draft.md
timber mango
I added to error checks to a script and MzPyLint is tagging me for lines too long. I can not move the error checks. Can I just disable the line too long check? If so, how?
timber mango
raise ValueError('Digit value must be an integer in the range: 0-3')
for bits in bitmasks:
if not 0 <= bits <= 0xFFFF:
raise ValueError('Bitmask value must be an integer in the range: 0-65535')
The error messages are the long lines. These are the only to places the script does not get past PyLint.
tulip sleet
if not 0 <= dig <= 3:
raise ValueError(
'Digit value must be an integer in the range: 0-3')
for bits in bitmasks:
if not 0 <= bits <= 0xFFFF:
raise ValueError(
'Bitmask value must be an integer in the range: 0-65535')
anything in parentheses can be split across multiple lines
You can also split a string with a \:
>>> print ("a\
... bc")
abc
the backslash-newline is ignored
timber mango
I hate doing that, but I guess I can.
tulip sleet
but i think the first thing is easier if it will work, just do a newline after the opening paren
timber mango
OK.
tulip sleet
and you can get pylint to ignore longer lines, but it makes for noise in the code, so if it can be made readable without, that's probably a bit better
timber mango
OK, I will do that then.
That should fix up my two PRs for the HT16K33 library. After that, I have one more thing to do to it.
timber mango
Out for now.
gusty topaz
Guys, I'm working on a project using Audio on Pyportal. Everything is working fine and as expected (I followed all the learn guides). When touching the screen a wav file plays per the code below. However, when the PP portal is idle (no interaction with it), I can hear a static/white noise coming out of the small speaker. Sometimes it gets really loud and then goes back to a low volume. Have any of you experienced something like this in the past?.
while True:
touch = ts.touch_point
# a larger time wait minimizes rogue coord. detection.
# 0.08 works fine.
time.sleep(0.08)
if touch:
audio.play(sound)
touch_data = get_data()
action(touch_data)
update_screen(touch_data)
@lapis hemlock yes the properties do work in circuitpython. ```>>> ulab.zeros(3).shape
(1, 3)
@onyx hinge I know what the problem is here, and it is really high on my list. As soon as properties are sorted out, I can fix this.
if it turns out that we have to make this different in circuitpython, then the code structure where all the "bindings" go in your ulab.c works great for us, we can put it in our
shared-bindings/ulab/__init__.c. but, is the payoff for having properties in circuitpython's ulab worth making code not portable between them? I'm less convinced. Would want to know what @slender iron thinks
@onyx hinge I am not going to break compatibility. In fact, I wanted to ask, whether you really have to move stuff into__init__.c. I could do that, if that is necessary. Properties are quite important, I think. I wouldn't like to remove them.
fallen anvil
On IMXRT UART pin definitions theres no pull enabled, and hysteresis is off. The lack of pull will generate noise and false inputs on the pin...was there a specific reason for doing this? Normally I would have 100K Pullup and hysteresis (because, why not?) on these. If there's no specific reason then I'll change it in the stuff I'm working on.
lapis hemlock
@onyx hinge @slender iron jimmo from micropython says that properties are currently not implemented in an accessible way in micropython, but that it could be made to work with some plumbing. He suggested that I lift the required code from circuitpython's runtime.c, so that it what I am going to do: https://forum.micropython.org/viewtopic.php?f=3&t=7762&p=44283#p44283
gusty topaz
Guys, I'm working on a project using Audio on Pyportal. Everything is working fine and as expected (I followed all the learn guides). When touching the screen a wav file plays per the code below. However, when the PP portal is idle (no interaction with it), I can hear a static/white noise coming out of the small speaker. Sometimes it gets really loud and then goes back to a low volume. Have any of you experienced something like this in the past?.
while True: touch = ts.touch_point # a larger time wait minimizes rogue coord. detection. # 0.08 works fine. time.sleep(0.08) if touch: audio.play(sound) touch_data = get_data() action(touch_data) update_screen(touch_data)
@gusty topaz has anybody seen this before?
timber mango
@gilded cradle I have been playing with a modified version of the animation script that works with auto_write=Trueon display initialization.
I switched the default for animate() to auto_write=False, which is what most of the animations I created require. This means fewer parameters have to be specified.
indigo wedge
I'm afk but can try to support
meager fog
im a little confused by what its askin here. i se the nxt MCU device, and the Boot Device to W25xxJV
then connected, and its sitting here with the bottom green bar advancing slowly
indigo wedge
Can you see the bootloader in devices in control panel?
indigo wedge
I think that's just like a fake loading bar,id you don't get the blue light then it didn't connect
indigo wedge
You can also try unchecking One Step and doing it one step at a time waiting a bit between
You can disable those in tools
Oh also I think I had to change mode in tools
To master
It was like 1st or 2nd submenu
meager fog
yah i changed to master
i get this in the cmd window
toolStatus: 0
commandOutput: {
"command" : "error-status",
"response" : [ 858993459 ],
"status" : {
"description" : "1450735702 (0x56787856) HAB disabled.",
"value" : 1450735702
}
}
Executing: C:\Users\ladyada\Dropbox (Personal)\micropython\NXP-MCUBootUtility\tools\sdphost\win\sdphost -t 50000 -u 0x1FC9,0x0145 -j -- error-status
toolStatus: 0
commandOutput: {
"command" : "error-status",
"response" : [ 858993459 ],
"status" : {
"description" : "1450735702 (0x56787856) HAB disabled.",
"value" : 1450735702
}
}
Executing: C:\Users\ladyada\Dropbox (Personal)\micropython\NXP-MCUBootUtility\tools\sdphost\win\sdphost -t 50000 -u 0x1FC9,0x0145 -j -- write-file 538990592 C:\Users\ladyada\Dropbox (Personal)\micropython\NXP-MCUBootUtility\src\targets\MIMXRT1011\ivt_flashloader.bin
toolStatus: 0
commandOutput: {
"command" : "write-file",
"response" : [ 2290649224 ],
"status" : {
"description" : "1450735702 (0x56787856) HAB disabled.",
"value" : 1450735702
}
}
Executing: C:\Users\ladyada\Dropbox (Personal)\micropython\NXP-MCUBootUtility\tools\sdphost\win\sdphost -t 50000 -u 0x1FC9,0x0145 -j -- jump-address 538990592
toolStatus: 0
commandOutput: {
"command" : "jump-address",
"response" : [],
"status" : {
"description" : "1450735702 (0x56787856) HAB disabled.",
"value" : 1450735702
}
}```but
Executing C:\Users\ladyada\Dropbox (Personal)\micropython\NXP-MCUBootUtility\tools\sdphost\win\sdphost -t 50000 -u 0x1FC9,0x0145 -j -- error-status
Executing C:\Users\ladyada\Dropbox (Personal)\micropython\NXP-MCUBootUtility\tools\sdphost\win\sdphost -t 50000 -u 0x1FC9,0x0145 -j -- error-status
Executing C:\Users\ladyada\Dropbox (Personal)\micropython\NXP-MCUBootUtility\tools\sdphost\win\sdphost -t 50000 -u 0x1FC9,0x0145 -j -- write-file 538990592 C:\Users\ladyada\Dropbox (Personal)\micropython\NXP-MCUBootUtility\src\targets\MIMXRT1011\ivt_flashloader.bin
Executing C:\Users\ladyada\Dropbox (Personal)\micropython\NXP-MCUBootUtility\tools\sdphost\win\sdphost -t 50000 -u 0x1FC9,0x0145 -j -- jump-address 538990592
so maybe it loads the bootloader but fails on the jump-address
it never comes back as the secondary flashloader
indigo wedge
Yes for me the initial bootloader disappears and a new one appears
Try disconnecting and connecting again sometimes its flimsy, I have to run for now
umbral dagger
@tulip sleet Thanks, Dan. Good call. Might have been nice to have a barrel jack as an alternate (higher current) power source.
meager fog
@indigo wedge did some light rework, checked values - 24 mhz crystal is giving me a nice 1Vpp but...still nothin
ill wait till ya send me boards
timber mango
I think a good use for the animation function I created is for small robot signaling and communication. Small robots do not have much space for networking and communication tools. I would love to expand the HT16K33 library to use multiple displays.
indigo wedge
@meager fog I think if you get the usb bootloader then the crystal should be fine
fallen anvil
@indigo wedge I'm implementing CTS/RTS/RS485 on IMXRT. That was straightforward, but I think I've found some issues with the serial which are down to the fsl driver rather than anything you've done...are you the right person to ping this off of?
indigo wedge
I'm preparing to go to sleep now but you can DM me and I'll read in the morning ๐
fallen anvil
OK. No worries.
I'll see if I can get closer to a resolution before sending you something.
cracked it - when asking for a uart read with timeout and the timeout triggered then the full requested length would be returned, even if no data had been received. That's because LPUART_TransferAboveReceive nukes the rxDataSize back to zero. Just changed the return to be self->handle.rxData-data and all is good for both the timeout and read complete cases. Sweet dreams.
manic glacierBOT
timber mango
Other than BDF, which font format is used throughout Circuitpython? There is no ยฐ (degree) symbol in font5x8.bin, which is used in the IS31FL3731 library for the Charlieplex displays. Would it be reasonable to convert the library to use a different font format? Would it be worthwhile to do this?
timber mango
I guess the "*" is better than nothing, but not as good as I would like.
timber mango
I just uploaded an alternate version of my ht16k33 animation demo to my git repo at https://github.com/geekguy-wy/Circuitpython_Goodies
onyx hinge
implemented github actions for ulab, but it needs some useful tests. https://github.com/v923z/micropython-ulab/pull/39
lapis hemlock
implemented github actions for ulab, but it needs some useful tests. https://github.com/v923z/micropython-ulab/pull/39
@onyx hinge Thanks Jeff, I have merged the PR.
pastel panther
heya @slender iron, if I get a seg fault trying to debug why a bootloader doesn't appear to be running properly is there something else I can look at to keep tracking down the problem? I'm about at the edge of my debugging ability:
ionic elk
I have gotten the flu real bad this weekend and might miss the start of next week, sorry! I'll check discord every day but probably won't be working on much since I'm riding it out at my folks house.
stuck elbow
get well soon
gusty topaz
Guys, I'm working on a project using Audio on Pyportal. Everything is working mostly fine (I followed all the learn guides). When touching the screen a wav file plays per the code below. However, when the PP portal is idle (no interaction with it), I can hear a static/white noise coming out of the small speaker. Sometimes it gets really loud and then goes back to a low volume. Have any of you experienced something like this in the past?. Here is the main part of the scrip (this is the only place where the audio.play() is called:
while True:
touch = ts.touch_point
# a larger time wait minimizes rogue coord. detection.
# 0.08 works fine.
time.sleep(0.08)
if touch:
audio.play(sound)
touch_data = get_data()
action(touch_data)
update_screen(touch_data)
Do you guys have any idea why the speaker would constantly generate that white noise?.
pastel panther
@ionic elk Dr's advice: Stay off the interwebs and drink plenty of fluids
bright aspen
Is there a way to generate an MCLK (audio system clock) for I2S output on Feather M4 Express or Feather STM32F405 Express?
main meteor
I know it works on the M4 Express https://learn.adafruit.com/adafruit-i2s-stereo-decoder-uda1334a/circuitpython-wiring-test
bright aspen
@main meteor Thanks! I have used the UDA1334 with the M4 Express. The uda1334a board doesn't need an an MCLK, it derives it from the BCLK. However, I'm now moving on to a DAC that needs an MCLK, a multiple of the LRCLK (WSEL) and locked in. The multiple is typically 256, I think 128n in general.
main meteor
Oho, you're already ahead of me then
onyx hinge
@bright aspen circuitpython doesn't do that right now afaik. Looks like the m4 hardware supports it so it should be possible to add.
Also hopefully the pin you need is broken out
lone sandal
bright aspen
@onyx hinge Thank you! I had been using the UDA1334 and another that did not need an MCLK, so I had forgotten about it and I'm in a panic. For my DAC the MCLK doesn't have to have a phase relationship with BCLK, it just needs to be at a frequency that is an appropriate multiple of LRCLK (eg, 12.2880 MHz or 18.4320 MHz) and derived from the same clock source (can't use a separate oscillator). Options that I'm currently considering are these:
2 Add MCLK to CircuitPython I2S for a candidate processor (as you suggested), using an all-the-pins for the prototype if need be.
3 Find another way to make a candidate processor generate, say, 12.2880 MHz.
4 Drop down to C.
5 Switch to a pic32mx270F256B (in a DIP).```@gilded cradle If you have a moment, can you review this? I tested it, it works. Extra stuff was from the original example and was unneeded. Would have made the explanation have unnecessary extra bits in it. https://github.com/adafruit/Adafruit_CircuitPython_CLUE/pull/17
gilded cradle
Sure @idle owl
gilded cradle
You're welcome
orchid basinBOT
pastel panther
heya @gilded cradle (or any other graphics savvy folks) take a look at this? A user is drawing arcs graphs using displayio and running into issues with rounding errors:
https://forums.adafruit.com/viewtopic.php?f=60&t=162185&p=798492#p798492
slender iron
@pastel panther that segfault looks a bit like a red herring to me
on arm you typically end up in the HardFault_Handler (you can set a breakpoint on it)
Just a reminder. The CircuitPython Weekly is tomorrow, Tuesday, at 11am Pacific / 2pm Eastern here on Discord. All are encouraged to attend. Meetings are normally Monday but today is a US holiday so we bumped it one day forwards. Meeting notes doc is here: https://docs.google.com/document/d/1cTRyH_oxjhgRwpQPM-oSOIB-chxbplCkyqx2NYRPJ8o/edit <@&356864093652516868>
slender iron
@meager fog working once sounds like it was loaded into memory
indigo wedge
@meager fog I can make a quick video of how I do it when I'm back from work
manic glacierBOT
indigo wedge
I'm now blocked waiting for PCBs for a week or two so will try to get more imx port work done ๐
manic glacierBOT
meager fog
@indigo wedge yah
so i did the pull-boot-pin low thing
the nxp loader at least got to the second stage - and i tried loading the feather_mimxrt1011-firmware.elf
but then nothing, and now i cant get back to the second stage
indigo wedge
hmm, did you configure the boot device before flashing?
fallen anvil
Welcome to the circle of hell that is imxrt boot configuration :-|
meager fog
if there's a step by step process - i'll need to know the steps ๐
i "Looks like loading the .elf into Application Image File field in the Image Generation Sequence tab and then pressing "All-In-One Action" flashes the device successfully!"
indigo wedge
yes, seems i missed a one-time step cause it was one time ๐
i'll make a video after work ๐
meager fog
ok but now i cannot get back into the RAM loader
do i have to pull the other boot pin?
on my metro design i put two switches, maybe i should finish that design ๐
indigo wedge
if you pull that line down it should always boot into rom loader the other one is already tied low iirc
fallen anvil
@indigo wedge Whats to work on? I'm just finishing up the RS485/RTS/CTS thing (I needed RS485, the others were virtually free) and I'm eyeIng CAN next, but I don't know if there's a standard python API for it?
indigo wedge
don't think so
there's a few things to do https://github.com/adafruit/circuitpython/labels/mimxrt10xx
also i need to add more functionality to my uf2 lib
fallen anvil
yeah, I need uf2 for 1021...that's quite high priority for me actually.
indigo wedge
would be cool if someone wanted to port that to a different manufacturer so we could figure out all the issues there
like nrf, stm or samd
fallen anvil
nrf is on my todo list if no-one else gets to it first, but it's well down that list
indigo wedge
my 1011 evk's USB connector broke off so i couldn't test on it anymore
meager fog
@indigo wedge ๐คท i can't get nxp bootloader to get back to the ram loader anymore
indigo wedge
you mean the stage 1 "SE" something something
strange cause that would mean it wrote something to the flash that makes the rom loader think that there is a valid fw there so it jumps but pulling that pin low should force rom loader
meager fog
yah
right
can you check the other boot pin
like are you sure both should not be pulled low
indigo wedge
ok so
sdcard d2 is boot_mode[0]
sdcard d1 is boot_mode[1]
both have internal pull downs, [1] has external pull up
so default is [1:0] = 10 - internal boot
if you put 0 on that resistor then we do [1:0] = 00 - boot from fuses which makes the rom loader check flash and boot if it finds something there
if you put 0 on the resistor and 3.3v on d2 of sdcard then we get [1:0] = 01 - force rom loader
so you could try that last thing, i understand you might need 4 hands for that ๐
indigo wedge
no, that's for the detect, you have to connect on the sdcard connector pad
might need test points for that next time but i didn't think they would be needed
indigo wedge
i could show on the render but no kicad at work
fallen anvil
@indigo wedge I've had the problem when you've managed to get some configuration into the flash which is sort-of-valid-but-not-quite. In that circumstance the easiest way out of it I've found was to stop the flash getting read sucessfully by just shorting one of its pins to ground on power-on. With the leaded chip that can easily be done by shorting pins 1 & 2.... (probably not being helpful here, I'll go back under my rock)
indigo wedge
guess that's an option @meager fog, short d1 and d2 or something like that when you connect usb, if the rom loaded pops up then you can remove the short and pull the resistor down as i said before
meager fog
@indigo wedge yah please show on render
fallen anvil
more dumb question time; my PR just failed CI because of check-translate. Am I supposed to submit locale/*.po and locale/circuitpython.pot as part of my patch, or did I miss something?
timber mango
@gilded cradle Modify the brightness property is PR #62.
timber mango
@gilded cradle Are my PRs OK then?
meager fog
@indigo wedge (i didnt solder on the SD card so it shouldnt be too hard to add jumper cables)
timber mango
There should be a #help-with-github thing.
I do not understand what to do to rebase PR #63.
haughty drift
Arturo182: Are you aware that one of these discs you show in you post was actually shown in an episode of the detective series Bosch (an Amazon video series)? They called it a tracking device.....
tidal kiln
slender iron
@fallen anvil yeah, when you add strings in translate() you also need to run make translate at the top level to update the files in locale
fallen anvil
OK, so I just commit the resulting files then?
done and pushed, let's see if we get a clean CI build...
slender iron
@fallen anvil yup! thanks!
manic glacierBOT
fallen anvil
@slender iron far be it for me to tell you how to run your project, but what is the idea behind committing derived files? Seems a bit unusual.
slender iron
it's so that translators don't need to run any tools, they can just edit the files
for example, I know at least one of our regular translators just edits the .po through the github ui
half sedge
Hi, I was checking for a Circuit Python board with both BLE and SD card. The idea was to record Heart Rate values (and/or debug information) to know the long time stability. And I did not find any. What would you add/do? (1) Log in the flash (never did that, but if not connected as storage, it should work)? (2) I have a 128x128 TFT display wiht uSD card, that I could use (would that work without using the display part of it)? (3) Is there a feather wing that would do that?
slender iron
@half sedge logging to sd card should work. there is someone on the adafruit support forum doing similar and they are having trouble logging at the data rate they'd hoped. You can definitely try a display with SD card slot for prototyping
I'm not sure if we have a featherwing with an sd card slot only
half sedge
From memory there is a loggin feather, but that would be M0 + SD.
timber mango
@half sedge You could add the AdaLogger Featherwing. It has an SD card and RTC. I am using that to timestamp and log data from a group of sensors. It works great!
slender iron
heh, I fail in remembering all featherwings ๐
timber mango
@slender iron That is why there is a great Community Brain here. ๐ ๐
I only remember the ones I am actually using. ๐
slender iron
๐
timber mango
That AdaLogger Featherwing is one of my favorites, and it is very inexpensive.
onyx hinge
@lapis hemlock do you intend for your "2d" branch to be merged to master at some point? What things need to happen before that can be done?
I moved Drew Fustini's hug reports to alphabetical order. I'll move status updates too, if it's out of sequence.
lapis hemlock
@onyx hinge I think, I review from you would be enough. If you think that it is OK for now, we could merge.
onyx hinge
(maybe he's going alphabetical by last name? hm)
@lapis hemlock I'd be happy to do that review. I have a couple of hours today I can work on stuff for adafruit, and anything that helps clear the path for our ulab pr in circuitpython would be good use of that time.
lapis hemlock
I have started to work on an improvement (matrix branch, not yet uploaded), but as I said earlier, the changes there are going to be internal.
onyx hinge
fwiw I also want to review some of the warnings that would occur during a circuitpython build if we used the "usual" C compiler flags, and fix them if it looks like they could cause a real issue or as long as the fixes don't dirty the code..
lapis hemlock
@onyx hinge I might go to bear early today. If, because of the time difference, you don't manage the review before that, you can merge into master, I don't mind.
@onyx hinge I can't comment on that, since I don't get any warnings in micropython.
onyx hinge
sure thing, with your permission
lapis hemlock
Oh, one thing: you wanted to pass ndarrays to DMA. You could do that by implementing the data property: https://docs.scipy.org/doc/numpy-1.13.0/reference/generated/numpy.ndarray.data.html#numpy.ndarray.data . Do you want to do that, or should I? It is very simple, one needs the old asbytearray code, and simply return the pointer to the array.
@onyx hinge this is ridiculous: I wanted to say that I might go to bed early. Ha!
onyx hinge
I guessed it
slender iron
@lapis hemlock make sure and read the hug reports for you in the circuitpython meeting notes if you are unable to make it ๐ https://docs.google.com/document/d/1cTRyH_oxjhgRwpQPM-oSOIB-chxbplCkyqx2NYRPJ8o/edit#
onyx hinge
Adding the data property sounds like a good idea. but in circuitpython with the "buffer protocol" implemented, it also works directly, like this in python3. ```>>> n = numpy.zeros(3)
memoryview(n)
<memory at 0x7f3bf590a588>
"hug reports" are when we thank people in the community for the good stuff they are doing
lapis hemlock
@onyx hinge As you wish, though, array.data is probably more numpythonic. In the long run I would like to have it, but it is not urgent.
"hug reports" are when we thank people in the community for the good stuff they are doing
@onyx hinge OK, I have learnt something new today. Thank!
slender iron
๐
timber mango
TEXT only.
lapis hemlock
I hope this is not too cheesy, but I have found the circuitpython community very friendly. Not only towards my own self, but I actually skim most of the messages here, and they are all extremely positive. I don't know who the real architect behind all this is, but they should definitely be proud of themselves. I can't place a hug/thumbs-up/high-five whatever here, because @timber mango requested TEXT only, so you have to put up with TEXT only:)
idle owl
@lapis hemlock It's not cheesy at all. That's the goal. We strive for that.
lapis hemlock
@idle owl very well done.
idle owl
CircuitPython Weekly: Construction-on-Kattni's-End-Edition. ๐
slender iron
<@&356864093652516868> meeting is starting
wraith tiger
I'm just lurking, btw.
timber mango
lurking today
graceful heart
lurking
marble hornet
Lurking
half sedge
Lurking
onyx hinge
@lapis hemlock join us in the voice chat if you can!
lapis hemlock
@onyx hinge I think I've got to take off now, but next time I would be happy to attend.
onyx hinge
see you ! It's usually mondays at this hour.
lapis hemlock
@onyx hinge so it was postponed because of the holiday.
onyx hinge
right, it was a US holiday yesterday.
lapis hemlock
OK, then till Monday!
tidal kiln
** lurking **
slender iron
ivory yew
Can't make it today ๐ญ but I am working on a new m0 based Eurorack module. ๐
onyx hinge
@ivory yew take care! ๐
idle owl
https://twitter.com/WillingCarol/status/1228698325693583360
https://anweshadas.in/february-pyladies-pune-workshop/
Twitter
tiny oriole
absolutely!
marble hornet
** text only
raven canopy
Gotta drop. ๐ฆ ๐
pastel panther
byee @raven canopy
onyx hinge
dan, curious whether the spi transaction length change made a noticible difference
tulip sleet
i could not tell! I was testing with turtle
onyx hinge
oh well.
tulip sleet
i needed some example that sent a lot more data
but it also provoked me to rewrite the DMA loops to be simpler
onyx hinge
cool beans
pastel panther
check this out!
raven canopy
yay! That meeting was cancelled. Now I get an extra 30 minutes before the next one. ๐
slender iron
@raven canopy great! you are next ๐
raven canopy
Yep
pastel panther
The arc graph thread on the forums
https://forums.adafruit.com/viewtopic.php?f=60&t=162185
onyx hinge
[oops mubes was status updates only]
idle owl
There's no hug report from mubes. Only Status Updates.
onyx hinge
if you didn't know what black is, https://black.readthedocs.io/en/stable/
half sedge
I will add in the note a Hug Report to @lone axle and @gilded cradle , I did not notice they added the background color to Label. I created that "issue", and only figured in this meeting that maybe that's what you were talking about.
timber mango
I have a hug report for @gilded cradle for coaching me through the PRs to the HT16K33 library stuff.
idle owl
@inland tusk Please mute.
timber mango
Working on the HT16K33 stuff has been so much fun!
lone axle
Seeing your messages go by in the chat inspired me to pick up one of the 8x8 grid feather wings over the weekend to play with ๐
half sedge
No thermal camera required if you have the Titano, I have code that do not require the cam.
onyx hinge
David Glaude is a secret time lord with two hearts
half sedge
@onyx hinge or a Klingon?
onyx hinge
@half sedge I did not know that
half sedge
@onyx hinge I just checked, it is a misconception, they have one heart, but with 8 chambers, where human have 4 chambers. Sorry for misleading. ๐
marble hornet
Ci stands for?
idle owl
Continuous Integration
marble hornet
Thanks
half sedge
@idle owl At one time I wanted PyPortal for PyGamer+AirLift... because I did not have a PyPortal yet.
tiny oriole
interested in helping out with demos once i get my new screen installed ๐
half sedge
@idle owl Maybe you can consider PyBadge for PyPortal, I did wear a PyPortal as a badge at Fosdem.
tiny oriole
awesome, sure
half sedge
I want to do BLE broadcast of the color it read.
Compatible with the Hide and Seek demo and all the Cheerlight I did.
lone axle
Neat idea. It's a bit of a different take on the Gizmo color sender / transmitter
allows for a wider group to use it too since it's not just the closest one only
half sedge
I wanted to do colour mixing using the Hide and Seek demo. You get the color, and you get the RSSI and you blend them to have the color of your "badge".
lone axle
I'm kicking around the thought of a "playground" type Mobile app kind of like the Circuit Playground Bluefruit one that is made to work with CLUE
marble hornet
So it would be device specific if rude?
half sedge
So if you are near Red and Blue, your badge get a mix of those. To choose your color, you scan a color with the CLUE reader.
onyx hinge
I love the quilt idea fwiw
prime flower
dropping out - heading out for lecture, listening later on spotify ๐
marble hornet
What about a mesh to find friends throughout pycon?
raven canopy
2nd day_job meeting cancelled. Best. Tuesday. Ever. (For this week, at least.)
idle owl
Another idea is like the interaction badge colors that grew out of autism conferences, where you have green if you want to talk to anyone, yellow if you only want to talk to friends, and red if you don't want to talk to anyone. So you could transmit green to others who are in a "green" mode to find new people to talk to.
So your CLUE does something if you're in a green mode and someone else in green mode is near enough to you.
slender iron
@idle owl maybe it can be scoped by topic?
idle owl
@slender iron Nice
onyx hinge
[dang sommersoft does a lot of stuff]
raven canopy
Not really. Just long-winded. ๐
tiny oriole
@idle owl That could be interesting. Maybe use A/B buttons to set your mood, and if it is in green mode, then maybe use the speaker to play a noise on both clues if someone else is green and near
idle owl
@tiny oriole Yeah, and to add Scott's suggestion, includes a series of topics that you can choose to enable/disable, and if someone else wants to discuss the same thing or "General discussion", then it connects.
inland tusk
I have a question about a buffer overflow when I tried to download a config over midi sysex.
timber mango
@lone axle You should get one of the HT16K33 Featherwings with the alphanumeric display. Making these do animations has been a blast!๐
graceful heart
might be worth looking into Yeoman or some other project generator
marble hornet
Could the continuous integration pull the repos when it sees said repoโs cookie template is out of date ?
idle owl
There's a lot of risk in that, unfortunately.
graceful heart
then you might be able to just do a yo upgrade automatically on each repo, assuming its a non breaking change
or manually to do it safely
raven canopy
That was part of the reason I broke out adabot's validators into a class. It's easier to add them.
marble hornet
What is โblack โ?
half sedge
a color?
onyx hinge
The uncompromising code formatter
By using Black, you agree to cede control over minutiae of hand-formatting. In return, Black gives you speed, determinism, and freedom from pycodestyle nagging about formatting. You will save time and mental energy for more important matters.
Black makes code review faster by producing the smallest diffs possible. Blackened code looks the same regardless of the project youโre reading. Formatting becomes transparent after a while and you can focus on the content instead.
timber mango
I soooooo want a CLUE, Feather Bluefruit Sense (does this have the 2MB for Circuitpython?), and a Circuitplayground Bluefruit!! I would pair the Sense and/or Circuitplayground Bluefruit with a CRICKIT for two robots. I already have both CRICKITs. I would also LOVE to help test the BLE stuff!
raven canopy
Caveat: it's one more tool to maintain... ๐
fallen anvil
@idle owl @onyx hinge sorry, did I do something wrong? New to this community, so nudges are welcome if I should do something different...
idle owl
@fallen anvil Not at all! We simply didn't want to miss your notes, Jeff didn't realise you'd only put notes in one place, not two.
half sedge
None mean you did not provide a parameter...
fallen anvil
Ah, fair enough...just wanted folks to know what I was up to in case it needs tuning or something.
pastel panther
idle owl
@fallen anvil Thank you for the update! We appreciate it. No worries, you're doing well ๐
marble hornet
Processing has great docs too. Just a though
lone axle
I think it would be great if the Learn Guide(s) / Github Readme / Readthedocs page for related modules and devices would all link to each other so you can get to any one of them from the other.
idle owl
@lone axle We try to do that with RTD and Learn Guides.
inland tusk
None usually means that a parameter has not been inialized or is not applicable.
raven canopy
@pastel panther so move from Sphinx to something else? rST has its limitations...
timber mango
I want to pair my PyGamer with wireless to make a robot controller.
Will there ever be a BLE Featherwing that can be added to boards that do not have it already?
marble hornet
Please post, would love to listen to it
half sedge
Today, for a board/sensor, you have: (1) A library (2) A learn guide (3) A simple_example (4) Readme.md (5) Doc generated from the code comment.
If you don't know, you could miss that there is (2) or (3).
marble hornet
Yes! I do
raven canopy
I mean, for nice expandable documentation, you'll have to move.
errant grail
Simple documentation mgt characteristics: identifying the owner and building in an expiration date.
slender iron
raven canopy
I do not disagree. At all.
half sedge
Docs in the code. Does that exist for Python?
idle owl
We already do that for libs.
And CP.
Some of the libraries are well documented within the code. Translates to Read the Docs.
half sedge
Could do pseudo code, just the function, not the content and add the doc to that for the code.
So write what the python description of the function would look like if they were Python (where they are C).
Then process the same way you do for library.
Cool.
slender iron
half sedge
A 5.1 with uLab is good for marketting uLab.
onyx hinge
+1
marble hornet
Editing GUI?
pastel panther
heading out, ttyl everyone ๐
idle owl
Need to drop off. Thanks everyone!
slender iron
timber mango
My muli-sensor script is almost 1000 lines of Circuitpython now, with quite a bit of comments. It is mostly code though.
slender iron
timber mango
I got an install deps failure on PR #65. How can I clear this?
Is that the one I need to rebase? How? I do not understand rebasing or how to use it yet.
idle owl
@slender iron @tulip sleet Is there a way to set a unique and chosen identifier when transmitting BLE? So if you and I are near each other, how do you show up to my board? Is there a way to set that to say "tannewt" or something so then I could have code that says "if tannewt, do foo"?
tulip sleet
there is an advertising name which you can set: BLERadio.name. As for unique, the serial number burned into the flash is available.
idle owl
@tulip sleet Both of those are perfect. Thanks
tulip sleet
I use the last four hex digitis of the uid for "CIRCUITPYxxxx"
raven canopy
@timber mango that failure looks Action's internal infrastructure related. I would just click the re-run button on it.
slender iron
@idle owl we can add something simple for it too
I love the idea of twitter handle because I always have a problem meeting folks I only know online
half sedge
Just an idea (that can be in the Weed). When there is an example in the library (simple text) write as a comment in the file, the original name of the file, or else link to GitHub. The idea is that the first thing you will do is copy that to your board and rename it to main.py. So keeping the name in help me.
Or maybe I do it wrong. I should keep the original name, and use main.py to import it.
onyx hinge
@slender iron just want to run something by you about packages implemented in C that affects how we are splitting ulab. I think this is inescapable, and related to how built in module dicts are read-only. Here's a session to show the oddities: ```>>> import ulab
dir(ulab)
['class', 'name', 'version', 'array', 'fft', 'filter', 'float', 'int16', 'int8', 'linalg', 'numerical', 'poly', 'uint16', 'uint8', 'vector']
ulab.fft.fft
<function>
import ulab.fft as fft
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
ImportError: no module named 'ulab'
from ulab import fft
@slender iron vs using ulab, ulab_fft, etc. Maybe we should go the latter way with underscores and not dots? because some of the stuff in that session just seems weird..
slender iron
@onyx hinge I think that is a micropython bug. try something similar in cpython. we should just fix it
@half sedge jerryn uses imports regularly instead of code.py
onyx hinge
I agree, it's different in cpython and even different in micropython when your package is made of py or mpy files
slender iron
@half sedge I like the idea of including the original name in a comment
idle owl
@slender iron I'm thinking for my interactive badge thing, you can add your friends to a list, and then set it to yellow mode, and it'll only react when your friends are near you. That involves setting what you're transmitting though. Which is why I asked
onyx hinge
I suspect the trouble of fixing it may be big. you need ulab's dict to have fft only after you import ulab.fft, but not before; but built in module dicts are rom'd and therefore readonly
slender iron
๐ it's important to note that anything you transmit is public
idle owl
right.
timber mango
I do not see a rerun button for this PR I need to restart.
half sedge
@slender iron I start to do that in my demo, when I had so many files around for various demo. I was doing that, and renaming main.demo_ble_sender.py
So it was clear this should be a main.py, but it is not because I work on another demo. So parking a demo on the device ready to use.
However Circuit Python and/or my computer was not super happy to see me rename the file, that were maybe open.
So change detection could fail and still run the old main.py.
I don't know if I am clear, but I did horrible thing that confused Circuit Python and/or my Computer.
Even Mu was sometime confused (or I was).
slender iron
ya, renaming the file can confuse editors. I run into that myself
pastel panther
@slender iron or other salty dogs:
is there a name or shorthand for this type of refactoring?
for number in range(0,10):
if number > 5:
if number in range(7, 20):
do_thing()
vs
for number in range(0, 10):
if number <= 5:
continue
if not number in range(7, 20):
continue
do_thing()
(moving to the later)
half sedge
What is the target date for demo code for the CLUE? Is that for the conference?
idle owl
@half sedge You're welcome to submit it at any time. To have it available for the conference, you'd want to submit it before the conference, yes, which is in the middle of April.
timber mango
@pastel panther I do not like the second one at all!
pastel panther
why?
slender iron
@pastel panther I kind of think of it as short circuiting
pastel panther
I was thinking "fail early"
idle owl
Failing early is ideal.
slender iron
I definitely prefer that style
half sedge
@idle owl Ok, then I keep that for later. But you are looking for something that work with zillions on CLUE at a conference...
idle owl
Not necessarily. It could work between two CLUEs and it's still something that people can use.
half sedge
@idle owl If I write something, I will test with CLUE talking to CPB+Gizmo (both have the same screen, A and B button, maybe more). I count on some of you to give it a try with more than one.
timber mango
I would do python for number in range(0, 10): if number > 5 and number in range(7, 20): do_thing()
idle owl
@half sedge That's fine, but please make it clear by noting in anything you submit that you have not tested it on multiple CLUE boards.
pastel panther
@timber mango sure this was a toy example to demonstrate the fail early type of refactoring
@timber mango I think your example is good when the conditions are functionally or otherwise related
timber mango
@pastel panther Ah, OK, well,I do not like it.
pastel panther
@timber mango ok, but why don't you like it
If you're onto something I'd like to learn from you
half sedge
@idle owl Did you find a simple/easy way to know on what board you are running? Just like your PyBadge code running on CLUE, how do you know you are on a CLUE?
@idle owl Looking for a "import board" + something
slender iron
@pastel panther do you have a real example you can link to?
pastel panther
probably, let me see
timber mango
@pastel panther I just cringe at the use of continue.
slender iron
why?
pastel panther
ok, why?
timber mango
I just do not think it has a place in that context.
Conditions should be together if they apply to the same thing.
timber mango
If that is the case, I might do:python if number > 5: if number in range(7, 20): do_thing()
half sedge
@idle owl I did not re-order the meeting notes.
idle owl
@slender iron ^^
pastel panther
@timber mango that' nesting is exactly what I'd be trying to avoid
timber mango
Why?
timber mango
I am not sure I agree with that. I have never had problems reading code with nested conditionals.
But then, that is my style of coding so I am naturally used to reading them.
slender iron
@half sedge @idle owl I did ๐
pastel panther
it also makes for longer lines than are needed and for me at least has a mental burden of having to track which branch you're on
timber mango
It can definitely lead to longer lines, and that has gotten me into a little trouble at least once.
pastel panther
sure, it's not a absolute but it's a possibility
for me it's more succinct to make the fact that you're stopping in certain conditions more explicit.
In the nested case you're "remembering" all the checks that you passed, to no benefit as far as I can tell.
timber mango
Even though I might cringe at the thought of using a particular style, I am flexible and can conform if required. Only if absolutely required though. ๐
pastel panther
We're not there yet
timber mango
I am soooooooooooooooooooo glad!
In the case of Adafruit, I would conform without reservations if it were needed.
pastel panther
@timber mango what do you think about my comment above about "remembering" all the checks that you've passed?
pastel panther
You can think of every indent as a piece of state. Why track it if it's not useful?
timber mango
I do not really know. I have never thought about it or had to think about it.
Just think about it as my native style and culture, so I do not have to think of those things because they are automatic for me.
timber mango
I feel like an old phaht with a young mind sometimes. I can usually soak up new information and apply it pretty quickly, even as I am still learning.
I do need good examples that actually work though. Much, if not most, of my Circuitpython stuff starts out by taking Adafruit Learn examples apart, analyzing them, and expanding them to what I need.
pastel panther
That's how I learn most new things: find working examples and modify them
I've tried and switched away from the style you prefer. I'll suggest that you maybe try the style I like for a while to see if you come around to it
or not, your choice
I say I "tried" your style but that's not right. It's how I wrote code for the first 5-10 years
timber mango
For instance, take the Arduino example for the Mini Robot Rover Chassis Kit. I am still learning from that code and have overloaded all the buttons and the d-pad of the app to get expanded functionality. I am going to recode that in Circuitpython.
@pastel panther I might do that the next time I start a project, which is often, and when I can use a new style from the start.
pastel panther
ya, I wouldn't switch mid project
timber mango
If Adafruit looked at my version of their Arduino code, they would probably not recognize much of it. ๐
pastel panther
ยฏ_(ใ)_/ยฏ
timber mango
Yeah, that!
I need to wire up a custom power cable because I can not find a pre-made one anywhere. ๐ฆ
I wonder if the 4400 maH Lion batter would fit on the three-deck round robot chassis. I know the 6600maH battery is too big.
fallen anvil
@slender iron Thanks for the review. Will there be an auto-tool for fixing code layout issues? I'm lazy...
slender iron
@fallen anvil yup! that is definitely a requirement
fallen anvil
Trouble is, the longer you wait, the more it mangles existing code
slender iron
looks like it will be astyle
ya, true
that's why I'm trying to keep the effort moving
fallen anvil
don't cae what it is, provided there's a config file :-) On some projects I use astyle, some uncrustify
slender iron
it's also a style that matches most of the existing code
looks up uncrustify
most of my experience is with clang-format
fallen anvil
Is there a 'beta'? I don't really want to deliberately submit mangled code
slender iron
ya, take a look at the issue I linked to
timber mango
@slender iron is http://astyle.sourceforge.net/ the one?
fallen anvil
yeah, I've used that too. It really doesn't matter, as long as there's one and 'make pretty' is a valid target :-)
slender iron
it has two similar astyle configs
@timber mango yup! related issue is https://github.com/micropython/micropython/issues/4223
@fallen anvil a make pretty target would be great
fallen anvil
The other nice thing about formatters, which people overlook, is I can use the format that I 'think' in, and then submit it in the standard format without upsetting anyone
The whole repository will need a run through with the tool though at some point before it can become 'standard'
slender iron
๐
ya, I'll do that once we decide
which is why I'm ok merging stuff that doesn't fit it now
timber mango
That site does not mention Python.
fallen anvil
I understand, but I won't do it deliberately. Lets get a tool in use, and deal with the pain it causes while it's still limited.
(Now I've got RS485 my modbus stuff will work, woohoo)
slender iron
@timber mango it is for formatting the C code in CircuitPython
@fallen anvil cool cool! thanks for being receptive to it
for python: https://github.com/psf/black
timber mango
Oh, OK. Will there be some format that will apply to Circuitpython also?
slender iron
@timber mango yup, see Black above
timber mango
I just installed it. ๐
I am still looking for the button to rerun checks on a PR but I can not find it.
idle owl
@timber mango You may not have permissions. Please post a link and I will rerun the job.
idle owl
Ok job running.
timber mango
Thanks
slender iron
Hi all, next week's CircuitPython Weekly meeting is the normal day and time, Monday the 24th at 11am Pacific / 2 pm Eastern, here on Discord. Everyone is encouraged to attend and participate. The notes doc is here: https://docs.google.com/document/d/1bGuche1SWmuFUABTWpuwapGnSaz4pG9yWR_Y0JOzdbw/edit?usp=sharing <@&356864093652516868>
timber mango
I have really had fun doing the animation and other stuff for the HT16K33 library. So cool!
timber mango
Puppy is feeling ignored. Got to go for a bit.
fallen anvil
@slender iron Hope I did the removal of #2618 properly! If so then it should be ready to ship now.
slender iron
k, will look now
@fallen anvil doesn't look like it worked. I think you want to merge adafruit/master in order to get the merge pr as well
it should also update the .po files
(merging the .po can be a pain)
fallen anvil
If I do that it gets upset about the po's. My git foo is weak. I guess I can just re-make those with make check-translate again?
slender iron
the easiest way is to checkout the versions from adafruit/master and then rerun make translate
fallen anvil
yeah, that's what I figured. Let's see what happens
slender iron
are you in the 'merging' state?
if git has put the diff lines in them then translate stuff will be unhappy
fallen anvil
I'm in the 'git-by-successive-approximation' state. cp -a is my friend.
slender iron
I have this git checkout --ours 5e789 -- locale/* in my shell history
fallen anvil
So I can just run that to get the locale back to a pre-state, then merge?
slender iron
you can be mid-merge when you do that I think
I'm not sure what the commit hash in there does
onyx hinge
it says which commit the content of "locale/*" is taken
but when combined with "--ours", it seems like one or the other must be ignored
slender iron
ya, could be my confusion
onyx hinge
I have my package style ulab working, but it is a bit of a jumble
(by using frozen modules)
@slender iron unless you have a way to get a toehold on fixing how packages made of built-in C modules work, I think this or "ulab_xxx" parallel modules are the way to go. If you have a direction to point me about the underlying problem I'm happy toc hase it down though
quitting for the day and won't look at this much before a week from now, so no rush to reply. just raising the issue
slender iron
@onyx hinge I haven't dug into it. I just like the idea of us acting more like real python.
raven canopy
watching that PyCon AU documentation talk...
@pastel panther mobile keyboard kept me from saying it earlier, but I would like a richer UI for docs. totally wanted some CSS collapsible boxes/dropdowns when working on the module support matrix. would've made navigating so much better than scroll-n-scan.
which, actually, pops something in my brain. @slender iron, thoughts on hooking the module support matrix into build_board_info.py and having that info available on circuitpython.org? have we discussed this (i can't remember)?
meager fog
@indigo wedge whew ok home nao!
raven canopy
๐ i'll drop it on my list to explore.
slender iron
@meager fog I'm around if you need someone to poke at the imx with
meager fog
@slender iron tough cause i have the imx feather
and artur is tryin' to help me get firmware loaded
slender iron
right, through the imx boot rom right?
fallen anvil
@slender iron @indigo wedge It's an interesting discussion. I don't think it does any harm for new developers to format according to the established norms of the project and it educates them in what's needed, but some things can, and should, be automated of course. It's no big deal, and I need something with my brain off having just done a few rounds with git.
meager fog
but its not sticking and i am unsure what bits to flip and the process to do so
@slender iron right - thru the nxp boot util which i think is win only
slender iron
@meager fog do you plan on using it when you flash them in production?
slender iron
I ran across the protocol at some point and we could hack up a loader python script
meager fog
@slender iron no
im looking to just get this feather booted so i can test it
then i design the test fixture later
slender iron
as a one off you could try flashing the spi flash directly
meager fog
ill use a jlink most likely
i could...
but i think its the fuses thats the issue, not the SPI flash
slender iron
or a jlink
meager fog
cause a programmed the SPI once
meager fog
yes
i mean, i can connect with gdb
i havent tried reading the fuses
also dont know what the fuse values should be
slender iron
can you see any activity on the spi lines? it should attempt to load a struct with info about the flash as a first step
meager fog
when does this first step occur?
slender iron
thanks for flexibility @fallen anvil
after reset I assume...
though it may only be on power up
slender iron
if there is no SPI activity then I'd believe the bootrom fuse bits are set wrong
slender iron
I believe the first read is always SPI
meager fog
i could look at clock line
slender iron
yup
slender iron
meager fog
with the default boot pin straps theres lots of activity on the SPI CS line
slender iron
ok, that is a good sign
that means the bootrom is trying to load from flash chip
what have you tried to load on it already?
meager fog
i see qspi traffic ---- except for the CLK pin
which seems to have some issue - not going full rail to rail
odd
meager fog
i see qspi traffic ---- except for the CLK pin
slender iron
I wonder if the first read is ok which would point to the config data structure being incorrect
fallen anvil
To 'become' qspi it has to have read properly. What's the b/w of your scope and how are you probing?
slender iron
ah, true
meager fog
i think the freq is wrong
my scope is only 100mhz so probably cant read/see the SCLK lines
fallen anvil
It should go to 133MHz when it goes qspi, so if you've not got decent earthing on your signal it won't look rail to rail
slender iron
what have you tried to load on it already?
meager fog
the uf2 bootloader
which loaded thru the romloader tool once...
now i cant get the feather to init thru the romloader
i have to look up the boot pin
slender iron
have you tried to load uf2 onto it?
slender iron
have you tried to load uf2 onto it?
meager fog
my scope is only 100mhz so probably cant read/see the SCLK lines
slender iron
hrm, ok. I haven't gotten that working myself.
slender iron
hrm, ok. I haven't gotten that working myself.
hrm, ok. I haven't gotten that working myself.
have you tried to load uf2 onto it?
fallen anvil
Thats me done for the night. Off to dream of git and merge trees.
slender iron
thanks @fallen anvil !
fallen anvil
Night. I've seen these boot issues before on IMXRT when the flash descriptor is vaguely corrupt. If you can get a JLink onto it then JLinkExe with exec EnableEraseAllFlashBanks followed by erase usually puts its brains back in. Good luck. Goodnight.
fallen anvil
Hmm...that should only take 10-20 secs.
Not too much burrito time. Anyway, from that raw state you'll now have access to at least the NXP usb loader...or just burn a hex using loadfile in JLinkExe. Don't use the bin cos you'll need to set the base address and I don't know if that bootable image patch has been folded in yet...
Right. Now it's really bedtime.
timber mango
@fallen anvil This is most likely true for those who are or have been developers. Those that are casual contributors here and there might not react the same way. I think we want to welcome and keep everyone. I have learned a thing or three from students at all levels.
I know that I was at first hesitant about joining in here even though I know Adafruit has a bunch of the best people to work with. I was still hesitant at jumping in and actually trying to change something. This is true even though I have been in and around computers since 7th grade and have been programming since then also. I have even had a couple paid programming jobs and I am still a little sensitive about some things.
All the people who work on the guts (low level C/C++) stuff are awesome. I have done a fair amount of C, but not enough C++ to claim it. I would not even consider trying to do anything with the guts of Circuitpython right now.
slender iron
@timber mango I'm always happy to help you (and anyone else) get ramped up on it.
timber mango
@slender iron Oh, I know! You have all been super to me ever since I got on the forums. Maybe if there are some easier issues, I might try poking a little deeper.