bb_paper inclusion in ESPHome | ESPHome | Page 1

steep quarry Feb 25, 2026, 9:15 AM

#

Hey @tropic musk I have made a thread here for your work implementing your epaper library into ESPHome.
Let us know how its going and if you have any questions here

#

cc @pseudo fossil

tropic musk Feb 25, 2026, 10:26 AM

#

thanks

tropic musk Feb 25, 2026, 2:15 PM

#

steep quarry cc <@1209965294492319815>

I'm having trouble with the codegen step. I'm getting some odd names coming out and I have no idea where the problem is 🙁

pseudo fossil Feb 25, 2026, 2:25 PM

#

Do you have a log? Is the code on gh?

tropic musk Feb 25, 2026, 5:42 PM

#

pseudo fossil Do you have a log? Is the code on gh?

It's currenty being rejected by github because of the ESPHome clang-format rule. I want to fix the bugs before I make it pretty.

#

The namespace/class/codegen step is completely opaque to me and it's generating strange results

#

src/main.cpp: In function 'void setup()':
src/main.cpp:603:55: error: invalid new-expression of abstract class type 'esphome::epaper::BB_EPaperBase'
  603 |   epaper_bb_epaperbase_id = new epaper::BB_EPaperBase();
      |                                                       ^
In file included from src/esphome.h:18,
                 from src/main.cpp:3:
src/esphome/components/bb_epaper/bb_epaperbase.h:11:7: note:   because the following virtual functions are pure within 'esphome::epaper::BB_EPaperBase':
   11 | class BB_EPaperBase : public display::DisplayBuffer,
      |       ^~~~~~~~~~~~~

#

my display.py:


BB_EPaperBase = epaper_ns.class_(
    "BB_EPaperBase", cg.PollingComponent, spi.SPIDevice, display.DisplayBuffer
)

CONFIG_SCHEMA = cv.Schema(
    display.FULL_DISPLAY_SCHEMA.extend(
        {
            cv.GenerateID(): cv.declare_id(BB_EPaperBase),
...

#

relevant part

#

I can create a new repo with just my component

#

I don't understand how the class name nesting gets so twisted up by codegen

#

I normally work in C, so this C++ syntax is quite confusing

#

it looks like something wrong with the way I'm extending the Display class

pseudo fossil Feb 25, 2026, 5:54 PM

#

tropic musk ```Generating LD script .pioenvs/esphome-web-1234/memory.ld src/main.cpp: In fun...

Did you override all the pure virtuals from display::DisplayBuffer? Ie something like:

   void draw_absolute_pixel_internal(int x, int y, Color color) override;
   int get_height_internal() override;
   int get_width_internal() override;
   display::DisplayType get_display_type() override { return display::DisplayType::DISPLAY_TYPE_BINARY; }

#

Or are any of those in the header but not implemented in the cpp?

raven isle Feb 25, 2026, 6:30 PM

#

tropic musk my display.py: ```epaper_ns = cg.esphome_ns.namespace("epaper") BB_EPaperBase =...

Your schema is defining the id as the base class, so when you generate the variable in the to_code, that's the class it tries to create which won't work.

#

I assume you have classes for each display?

tropic musk Feb 25, 2026, 7:08 PM

#

raven isle Your schema is defining the id as the base class, so when you generate the varia...

One class for all displays, each display gets an enumerated value

tropic musk Feb 25, 2026, 7:08 PM

#

raven isle Your schema is defining the id as the base class, so when you generate the varia...

I have no idea what ID and base class means in this sense

#

I thought it would be relatively simple to write a wrapper class over bb_epaper

#

I had to write some new features such as I/O through function pointers to get the I/O to work in ESPHome

#

I don't know how to "gen" the code to make this simple

#

I don't need components or subclasses

raven isle Feb 25, 2026, 7:10 PM

#

I wondered how well this was going to integrate when I first heard about it...

tropic musk Feb 25, 2026, 7:11 PM

#

should be quite straightforward if not for the class twist-up

#

I'll create a public repo for the code, hang on

#

https://github.com/bitbank2/esp_component

tropic musk Feb 25, 2026, 7:16 PM

#

raven isle Your schema is defining the id as the base class, so when you generate the varia...

Is there some easy explanation to connect the dots between the Python and the C++ that gets generated?

raven isle Feb 25, 2026, 7:17 PM

#

It's quite direct. It's the to_code that generates it.

tropic musk Feb 25, 2026, 7:17 PM

#

right, but Python->C++ syntax makes no sense

#

if it's just the missing virtual functions, then that should be easy to solve

raven isle Feb 25, 2026, 7:18 PM

#

var = cg.new_Pvariable(config[CONF_ID]) creates a new variable with the class defined in CONF_ID, so BB_EPaperBase in this case.

#

You can see the generated code in main.cpp.

tropic musk Feb 25, 2026, 7:18 PM

#

none of that syntax makes sense to me. What kind of programmer would work in that world ?

raven isle Feb 25, 2026, 7:19 PM

#

I don't see any issues with it... 🤷‍♂️

tropic musk Feb 25, 2026, 7:19 PM

#

so you're a Python programmer?

raven isle Feb 25, 2026, 7:19 PM

#

yes, and whatever other languages 🙂

tropic musk Feb 25, 2026, 7:19 PM

#

I never work in Python, and don't get into the weeds of C++, so it doesn't make sense to me

raven isle Feb 25, 2026, 7:21 PM

#

esphome gets into deep details of both.

tropic musk Feb 25, 2026, 7:21 PM

#

I'm less interested in becoming an expert in this subject and more interested in just getting this simple project working

#

my time is very limited and the frustration increases exponentially the more time I have to dedicate to this

pseudo fossil Feb 25, 2026, 7:23 PM

#

Are all the virtuals implemented? If your code is on GH it will be helpful.

tropic musk Feb 25, 2026, 7:24 PM

#

see link above

#

I'm adding the virtual functions now

pseudo fossil Feb 25, 2026, 7:25 PM

#

If pure virtuals are not implemented you will get that error when it tries to create the object in main.cpp

tropic musk Feb 25, 2026, 7:26 PM

#

fixing now...

#

I'm looking at the ST7789 implementation and it doesn't have a couple of the virtual functions that my build is complaining are missing

#

Thanks @pseudo fossil & @raven isle most errors are resolved now. I'll let you know if I have any more questions.

tropic musk Feb 26, 2026, 12:09 PM

#

Hi @raven isle closer, but still not working

#

my code all compiles without errors, but the display component is never called

#

  - platform: bb_epaper
    id: waveshare_esp32s3
    cs_pin: GPIO11
    dc_pin: GPIO10
    busy_pin: GPIO8
    reset_pin: GPIO9
    model: ep154_200x200
    lambda: |-
        it.print(0, 0, id(Roboto_10), "Hello World!");

#

I added logging to the main methods and they never get printed

#

Am I missing something obvious?

#

(WiFi succeeds)

#

[13:09:39.476][I][safe_mode:054]: Boot seems successful; resetting boot loop counter

pseudo fossil Feb 26, 2026, 12:26 PM

#

tropic musk I added logging to the main methods and they never get printed

Into setup? Are you using usb / serial? Setup is run before wifi connects so need to use usb.

tropic musk Feb 26, 2026, 12:26 PM

#

I'm using USB

#

I added logging to all main methods

#

nothing ever prints

#

initial log:

INFO Successfully uploaded program.
INFO Starting log output from /dev/cu.usbmodem113201 with baud rate 115200
[13:03:39.425][I][app:084]: Running through setup()
[13:03:39.426][V][app:085]: Sorting components by setup priority
[13:03:39.427][V][component:201]: Setup wifi
[13:03:39.444][C][wifi:373]: Starting

pseudo fossil Feb 26, 2026, 12:29 PM

#

Code in gh is up to date?

tropic musk Feb 26, 2026, 1:31 PM

#

not yet; I'll poke through the main.cpp to see if I can figure out the issue

steep quarry Feb 26, 2026, 7:03 PM

#

tropic musk not yet; I'll poke through the main.cpp to see if I can figure out the issue

I would suggest pushing changes, even in its broken state so we can view exactly the current state. it is extremely hard to help and debug without looking at the code and causes more delays

tropic musk Feb 26, 2026, 7:03 PM

#

steep quarry I would suggest pushing changes, even in its broken state so we can view exactly...

I will - I need to get it as far as I can first. It's starting to work.

tropic musk Feb 28, 2026, 2:27 PM

#

hi @raven isle I've made a lot of progress with the bb_epaper component, but there are still some issues to resolve.

#

The initialization and update methods seem to be working, but my yaml lamba doesn't result in any fill or pixel function calls

#

  - platform: bb_epaper
    update_interval: 30s
    id: waveshare_esp32s3
    cs_pin: GPIO11
    dc_pin: GPIO10
    busy_pin: GPIO8
    reset_pin: GPIO9
    model: ep154_200x200
    lambda: |-
       it.fill(id(Color::BLACK));
       it.print(20, 60, id(medium), "This is a test");

#

it compiles and runs without any errors. I put log print statements in the fill() and draw_absolute_pixel_internal() methods, but they're never called.

pseudo fossil Feb 28, 2026, 2:30 PM

#

Is your code up to date in gh?

tropic musk Feb 28, 2026, 2:30 PM

#

can you think of a reason that the fill and pixel methods wouldn't get called?

#

possible C++ override syntax issue?

#

I'll push the component source; hang on

#

ok @raven isle newest component code is here:

#

https://github.com/bitbank2/esp_component

pseudo fossil Feb 28, 2026, 2:33 PM

#

Very hard to debug something without seeing the code. Can always create a dev branch and push to that

tropic musk Feb 28, 2026, 2:34 PM

#

pseudo fossil Very hard to debug something without seeing the code. Can always create a dev br...

see above - code is now visible

pseudo fossil Feb 28, 2026, 2:36 PM

#

tropic musk see above - code is now visible

Looks like you are missing https://github.com/esphome/esphome/blob/dev/esphome/components/st7735/st7735.cpp#L294

tropic musk Feb 28, 2026, 2:36 PM

#

right now I'm getting successful updates, but the display is blank

tropic musk Feb 28, 2026, 2:37 PM

#

pseudo fossil Looks like you are missing https://github.com/esphome/esphome/blob/dev/esphome/c...

aren't those internal methods specific to that display component?

#

the call to update() from ESPHome is succeeding with my library

pseudo fossil Feb 28, 2026, 2:39 PM

#

do_update_ is some method from the display component, you inherit from DisplayBuffer which inherits from Display

tropic musk Feb 28, 2026, 2:40 PM

#

still don't follow you. It's not a problem with update

pseudo fossil Feb 28, 2026, 2:40 PM

#

Seems like you should be calling it but I havent done much display stuff

tropic musk Feb 28, 2026, 2:40 PM

#

trailing underscore in the name indicates it's private to that repo and not an override method

pseudo fossil Feb 28, 2026, 2:41 PM

#

It is protected. You inherited from display so that is part of your class

#

I am pretty sure that calls the lambda

#

Without it nothing will get rendered

tropic musk Feb 28, 2026, 2:41 PM

#

I think you have it backwards

pseudo fossil Feb 28, 2026, 2:41 PM

#

Did you try it?

tropic musk Feb 28, 2026, 2:42 PM

#

that method is protected because of the extra layers of hierarchy of that library

#

no need - update is getting called and succeeding in my library

#

update gets called every 30s as it should

#

that's not the issue

pseudo fossil Feb 28, 2026, 2:42 PM

#

Yes I know. Add this to your update function

tropic musk Feb 28, 2026, 2:42 PM

#

makes no sense because the fill and pixel functions are never called

#

they would need to be called BEFORE the update

#

even if it happens once

#

if you look in the waveshare display driver there is no update_() method

#

ok, I'll try

pseudo fossil Feb 28, 2026, 2:46 PM

#

tropic musk if you look in the waveshare display driver there is no update_() method

There is https://github.com/esphome/esphome/blob/dev/esphome/components/waveshare_epaper/waveshare_epaper.cpp#L171

tropic musk Feb 28, 2026, 2:46 PM

#

I see - hang on

pseudo fossil Feb 28, 2026, 2:47 PM

#

The lambda is called on this line https://github.com/esphome/esphome/blob/dev/esphome/components/display/display.cpp#L697 so without that call nothing is rendered

tropic musk Feb 28, 2026, 2:47 PM

#

ah - thanks for being patient

#

compiling...

#

makes sense - it allows the display component to skip a redraw if necessary

#

thanks @pseudo fossil - working now 🙂

#

tropic musk Feb 28, 2026, 4:22 PM

#

Now that it appears to be working, the bigger question is how we should present it to the users.

#

In it's current form, you specific the SPI connection and the panel type.

#

But... I've created a much simpler way for IoT devices with built in displays. For example, the Elecrow CrowPanel devices can be used with just a single line of code:
epaper.begin(EPD_CROWPANEL29);

#

The configuration is contained within bb_epaper, so the user doesn't have to fumble with GPIO connections

#

It currently supports 14 products like that

#

what do you think @raven isle @pseudo fossil ? I currently access SPI directly, but I could probably adapt it to use the ESPHome SPI component. Maybe I can create a second component for use with these products or a generic one for everything else.

tropic musk Feb 28, 2026, 6:29 PM

#

one last question...

#

to enable either the iot device name or raw panel, I need SPI to be optional for the display component. Is that possible?

#

@raven isle

#

when I use the iot device name, all hardware initialization and communication is handled by my library

#

I would like the user to not have to define the SPI bus

#

or will that require a new ESPHome component?

raven isle Feb 28, 2026, 8:32 PM

#

I think that's going to be a problem.

tropic musk Feb 28, 2026, 10:58 PM

#

I got it working

#

  - platform: bb_epaper_iot
    update_interval: 60s
    id: test_epd
    model: EPD_CROWPANEL29
    lambda: |-
       it.fill(id(Color::BLACK));
       it.print(0, 0, id(medium), "bb_epaper test");
       it.printf(0, 80, id(medium), "one line config!");

#

raven isle Mar 1, 2026, 12:13 AM

#

I mean I think it will be a problem getting it accepted.

tropic musk Mar 1, 2026, 1:05 PM

#

https://youtu.be/-B1YK6_o7dI

#

Hey @raven isle how do you like no-flicker e-paper for ESPHome? 🙂

#

I'm experimenting with a default update type of partial updates with a fast update every 16 updates to eliminate any possible ghosting

tropic musk Mar 2, 2026, 9:32 AM

#

@steep quarry I need some feedback on these decision. Should I expose the refresh type (full, fast, partial) or just do the least flickery way all the time?

#

(see youtube video above)

tropic musk Mar 2, 2026, 6:50 PM

#

Hi @raven isle I'm working on supporting color e-paper displays and noticed something odd. The Display class call to get_display_type() isn't called (I put a log print statement in it). Is this not really needed?

#

I'm thinking that I should just do a best match of the requested color.

pseudo fossil Mar 2, 2026, 7:05 PM

#

tropic musk Hi <@630900454926319635> I'm working on supporting color e-paper displays and no...

It is called in the test card code but needs to be implemented as it is pure virtual

tropic musk Mar 2, 2026, 7:07 PM

#

sorry, I don't follow you

#

for normal display usage will it get called?

pseudo fossil Mar 2, 2026, 7:13 PM

#

I believe if you do something like this it will be called:

  display:
    - platform: xxx
      show_test_card: true

tropic musk Mar 2, 2026, 7:29 PM

#

ok

steep quarry Mar 2, 2026, 8:12 PM

#

tropic musk <@205416954863812609> I need some feedback on these decision. Should I expose th...

I would expose them all

tropic musk Mar 8, 2026, 11:27 PM

#

final vault Mar 9, 2026, 2:11 AM

#

nice! what board is that?

tropic musk Mar 9, 2026, 8:38 AM

#

final vault nice! what board is that?

Waveshare ESP32-S3-ePaper-3.97"

steep quarry Mar 11, 2026, 6:08 PM

#

Hey @tropic musk Is the work in a state where it could be opened as a PR so we can see it with the ESPHome codebase?

tropic musk Mar 11, 2026, 6:09 PM

#

steep quarry Hey <@333625402272645122> Is the work in a state where it could be opened as a P...

It's curently in a separate repo that's easy to manually copy into the /components folder to test. Probably not ready to merge yet

#

https://github.com/bitbank2/esp_component

#

I'm rewriting my bb_temperature library to work with the ESPHome I2C component and then I'll add my bb_captouch (pretty much covers all common capacitive touch screens)

#

I would appreciate any and all feedback on the bb_epaper work so far.

#

It already is a major step up from the waveshare library because it allows you to more easily choose the type of panel and the update mode

#

I created a second component (bb_epaper_iot) that simplifies things further for supported ESP32+EPD products like the ones from Waveshare. However, it currently uses the Arduino or ESP-IDF SPI library directly instead of using the one in ESPHome.

steep quarry Mar 11, 2026, 7:09 PM

#

tropic musk It's curently in a separate repo that's easy to manually copy into the /componen...

Have you seen external_components?

tropic musk Mar 11, 2026, 7:47 PM

#

Yes and no ☺️. Are they written differently?

#

@steep quarry

raven isle Mar 11, 2026, 7:52 PM

#

It lets you supply components from an external source: https://esphome.io/components/external_components/

#

Exactly the same as the built-in ones.

steep quarry Mar 11, 2026, 7:56 PM

#

Just wasnt sure given:

easy to manually copy into the /components folder to test

tropic musk Mar 13, 2026, 4:36 PM

#

I'll redo the file structure to work as an external component

tropic musk Apr 11, 2026, 4:25 PM

#

@steep quarry I'm finally sitting down to work on making my component conform to the ESPHome coding standards. I've got clang-format and clang-tidy installed, but I'm not sure if I'm using them correctly. I tried running the scripts (clang-format and clang-tidy). Both give errors

#

The clang-format script says that I may not have it installed in the environment. It is installed, yet shows this when I pass --version as the only parameter:

zsh: /Users/laurencebank/Projects/esphome/venv/bin/clang-format: bad interpreter: /Users/laurencebank/Projects/esphome/venv/bin/python3: no such file or directory
clang-format version 22.1.3

#

Running the clang-tidy script gives these errors:

  File "/Users/laurencebank/Projects/esphome/esphome/components/bb_epaper/../../../script/clang-tidy", line 357, in <module>
    sys.exit(main())
  File "/Users/laurencebank/Projects/esphome/esphome/components/bb_epaper/../../../script/clang-tidy", line 285, in main
    idedata = load_idedata(args.environment)
  File "/Users/laurencebank/Projects/esphome/script/helpers.py", line 546, in load_idedata
    stdout = subprocess.check_output(
  File "/Library/Developer/CommandLineTools/Library/Frameworks/Python3.framework/Versions/3.9/lib/python3.9/subprocess.py", line 424, in check_output
    return run(*popenargs, stdout=PIPE, timeout=timeout, check=True,
  File "/Library/Developer/CommandLineTools/Library/Frameworks/Python3.framework/Versions/3.9/lib/python3.9/subprocess.py", line 505, in run
    with Popen(*popenargs, **kwargs) as process:
  File "/Library/Developer/CommandLineTools/Library/Frameworks/Python3.framework/Versions/3.9/lib/python3.9/subprocess.py", line 951, in __init__
    self._execute_child(args, executable, preexec_fn, close_fds,
  File "/Library/Developer/CommandLineTools/Library/Frameworks/Python3.framework/Versions/3.9/lib/python3.9/subprocess.py", line 1821, in _execute_child
    raise child_exception_type(errno_num, err_msg, err_filename)
FileNotFoundError: [Errno 2] No such file or directory: 'pio'

#

hmm - I just set up a new machine and it's missing some things. Let me work on this a bit more...

pseudo fossil Apr 11, 2026, 4:32 PM

#

script/clang-tidy --all-headers --environment esp32-arduino-tidy and script/clang-tidy --all-headers --environment esp32-idf-tidy --grep USE_ESP_IDF

#

have to do source venv/bin/activate first

tropic musk Apr 11, 2026, 4:33 PM

#

I did activate, but something else is broken

#

just running esphome from the command line gives this:

#

(venv) laurencebank@4a:5f:88:47:24:de esphome % esphome
zsh: /Users/laurencebank/Projects/esphome/venv/bin/esphome: bad interpreter: /Users/laurencebank/Projects/esphome/venv/bin/python3: no such file or directory
usage: esphome [-h] [-v] [-q] [-l {DEBUG,INFO,WARNING,ERROR,CRITICAL}]
[-s key value] [--mdns-address-cache MDNS_ADDRESS_CACHE]
[--dns-address-cache DNS_ADDRESS_CACHE] [--testing-mode]
[--version]
command ...

#

bad interpreter?

#

hang on, new machine problems - let me create a new venv

pseudo fossil Apr 11, 2026, 4:36 PM

#

script/setup
source venv/bin/activate
script/clang-tidy --all-headers --environment esp32-arduino-tidy
script/clang-tidy --all-headers --environment esp32-idf-tidy --grep USE_ESP_IDF

tropic musk Apr 11, 2026, 4:36 PM

#

it was that the new machine's python3 was in a different spot, but I had restored the esphome repo from the old machine

#

new python weirdness

#

(venv) laurencebank@4a:5f:88:47:24:de esphome % script/setup
Using Python 3.9.6 environment at: venv
Checked 2 packages in 11ms
Using Python 3.9.6 environment at: venv
× No solution found when resolving dependencies:
╰─▶ Because only esphome[dev]==2026.1.0.dev0 is available and the current
Python version (3.9.6) does not satisfy Python>=3.11.0,<3.14, we can
conclude that all versions of esphome[dev] cannot be used.
And because you require esphome[dev], we can conclude that your
requirements are unsatisfiable.

pseudo fossil Apr 11, 2026, 4:38 PM

#

Yeah 3.9 is too old need >=3.11

tropic musk Apr 11, 2026, 4:38 PM

#

strange, brew installed 3.14

#

how do I get the brew version to be seen in the esphome environment?

pseudo fossil Apr 11, 2026, 4:39 PM

#

I have never used brew

#

Why not do it natively?

tropic musk Apr 11, 2026, 4:39 PM

#

natively how?

pseudo fossil Apr 11, 2026, 4:40 PM

#

Clone the repo and just run the setup script?

tropic musk Apr 11, 2026, 4:40 PM

#

what repo?

pseudo fossil Apr 11, 2026, 4:40 PM

#

The esphome repo

tropic musk Apr 11, 2026, 4:40 PM

#

that has no bearing on the python3 problem

pseudo fossil Apr 11, 2026, 4:41 PM

#

I dont know what brew has to do with this but I have never used it.

tropic musk Apr 11, 2026, 4:41 PM

#

homebrew has python3 3.14 installed, but the system still references python 3.9.6

#

I need to fix the path order

#

@pseudo fossil thanks for your help. I had to switch the path order of zsh for it to find the correct Python version. Now things are working.

tropic musk Apr 11, 2026, 5:23 PM

#

@pseudo fossil closer... still getting errors from the clang-tidy script and parameters you provided:

#


Total cpp files to check: 1
Added all-include header file, new total: 2
Final files to process:
    /Users/laurencebank/Projects/esphome/.temp/all-include.cpp
    bb_epaperbase.cpp
Loading IDE data for environment 'esp32-arduino-tidy'...
NotPlatformIOProjectError: Not a PlatformIO project. `platformio.ini` file has not been found in current working directory (/Users/laurencebank/Projects/esphome/esphome/components/bb_epaper). To initialize new project please use `platformio project init` command

pseudo fossil Apr 11, 2026, 6:08 PM

#

platformio.ini is in the root dir: https://github.com/esphome/esphome/blob/dev/platformio.ini

#

run it from the root dir

tropic musk Apr 11, 2026, 6:23 PM

#

pseudo fossil platformio.ini is in the root dir: https://github.com/esphome/esphome/blob/dev/p...

sorry, can you clarify? I'm running it in the directory with the files to modify. How would the command line look for your idea?

#

This back and forth makes me think this would be a good topic to add to the developer documentation. I couldn't find any useful info there.

pseudo fossil Apr 11, 2026, 6:56 PM

#

You cant run it on a single component, run it in the root esphome folder (I gave the commands above)

#

Most people just submit a PR and have CI run it

#

script/setup
source venv/bin/activate
script/clang-tidy --all-headers --environment esp32-arduino-tidy
script/clang-tidy --all-headers --environment esp32-idf-tidy --grep USE_ESP_IDF

tropic musk Apr 11, 2026, 7:26 PM

#

pseudo fossil You cant run it on a single component, run it in the root esphome folder (I gave...

Can you describe the PR/CI process? When I tried to create a new branch with this code, Github blocked it because it didn't conform to the code style rules.

#

So you're saying to just use the error info from trying to create a PR to make the edits instead of letting clang-tidy do them for me?

pseudo fossil Apr 11, 2026, 7:30 PM

#

tropic musk Can you describe the PR/CI process? When I tried to create a new branch with thi...

Yes there is also a pre commit check.

#

You will need to fix those errors first

tropic musk Apr 11, 2026, 7:59 PM

#

pseudo fossil Yes there is also a pre commit check.

I'm still getting errors just running the script

#

  File "/Users/laurencebank/Projects/esphome/script/clang-tidy", line 357, in <module>
    sys.exit(main())
             ~~~~^^
  File "/Users/laurencebank/Projects/esphome/script/clang-tidy", line 285, in main
    idedata = load_idedata(args.environment)
  File "/Users/laurencebank/Projects/esphome/script/helpers.py", line 546, in load_idedata
    stdout = subprocess.check_output(
        ["pio", "run", "-t", "idedata", "-e", environment]
    )
  File "/opt/homebrew/Cellar/[email protected]/3.14.4/Frameworks/Python.framework/Versions/3.14/lib/python3.14/subprocess.py", line 473, in check_output
    return run(*popenargs, stdout=PIPE, timeout=timeout, check=True,
           ~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
               **kwargs).stdout
               ^^^^^^^^^
  File "/opt/homebrew/Cellar/[email protected]/3.14.4/Frameworks/Python.framework/Versions/3.14/lib/python3.14/subprocess.py", line 578, in run
    raise CalledProcessError(retcode, process.args,
                             output=stdout, stderr=stderr)
subprocess.CalledProcessError: Command '['pio', 'run', '-t', 'idedata', '-e', 'esp32-arduino-tidy']' returned non-zero exit status 1.

#

ah - new error: Loading IDE data for environment 'esp32-arduino-tidy'...
ERROR: Python version must be between 3.10 and 3.13.

#

The current version of Python on Homebrew is 3.14 and this seems to cause a problem for the script

pseudo fossil Apr 11, 2026, 8:01 PM

#

3.14 works unless you are on an old version

tropic musk Apr 11, 2026, 8:01 PM

#

old version of what?

pseudo fossil Apr 11, 2026, 8:01 PM

#

Or havent run the setup script recently

#

ESPHome

tropic musk Apr 11, 2026, 8:02 PM

#

I just did a git pull to get the latest. How can I tell

#

?

#

New machine - I just installed epshome

pseudo fossil Apr 11, 2026, 8:03 PM

#

I ran it today on 3.14 and it was fine

tropic musk Apr 11, 2026, 8:03 PM

#

This experience reinforces my belief that depending on Python for these tools creates a huge time expense for everyone 🙁

pseudo fossil Apr 11, 2026, 8:03 PM

#

Submit a draft pr it will run it

tropic musk Apr 11, 2026, 8:03 PM

#

I'd like to understand this roadblock

pseudo fossil Apr 11, 2026, 8:04 PM

#

Homebrew seems to mess things up, python is fine

tropic musk Apr 11, 2026, 8:04 PM

#

I'm using a newly installed esphome and the repo is up to date

#

I would assume the inverse

#

Python is always having trouble with version conflicts - it's a feature

pseudo fossil Apr 11, 2026, 8:04 PM

#

We have so many issues where people are using homebrew and can't build

tropic musk Apr 11, 2026, 8:05 PM

#

what's the "golden" OS to use for ESPHome? Linux?

pseudo fossil Apr 11, 2026, 8:05 PM

#

That is what I use. No brew. Just python

tropic musk Apr 11, 2026, 8:05 PM

#

homebrew is a package manager like apt, nothing to do with interfering in use of the tools

pseudo fossil Apr 11, 2026, 8:06 PM

#

Well people try to install ESPHome with it. It messes with the paths.

tropic musk Apr 11, 2026, 8:06 PM

#

I don't want to create extra work for you. It shouldn't take so much effort just to run clang-tidy

#

I can try this on my Linux laptop

pseudo fossil Apr 11, 2026, 8:07 PM

#

I gave the commands above. I tested them and they work.

tropic musk Apr 11, 2026, 8:08 PM

#

That's not helpful. If it's only designed for Linux, then ok.

pseudo fossil Apr 11, 2026, 8:08 PM

#

I don't have a mac to try it on

tropic musk Apr 11, 2026, 8:08 PM

#

Does anyone test it on MacOS or Windows?

pseudo fossil Apr 11, 2026, 8:09 PM

#

I know Nick runs on a mac, maybe Clyde does too. But most people just get CI to run it

#

Almost no one runs it locally

tropic musk Apr 11, 2026, 8:09 PM

#

the build system has worked fine so far on MacOS, it just seems to be an issue with running the clang-tidy script

#

when you say "get CI to run it", can you elaborate?

pseudo fossil Apr 11, 2026, 8:10 PM

#

By creating a PR

tropic musk Apr 11, 2026, 8:10 PM

#

my last try creating a PR went south. Lots of errors and a rejection. How is the work flow supposed to go?

pseudo fossil Apr 11, 2026, 8:11 PM

#

Yes there is a pre commit check before you can create one.

tropic musk Apr 11, 2026, 8:11 PM

#

right

pseudo fossil Apr 11, 2026, 8:11 PM

#

You have to fix those before you create a PR

tropic musk Apr 11, 2026, 8:11 PM

#

telling me "code rejected for formatting error" isn't actionable

pseudo fossil Apr 11, 2026, 8:11 PM

#

Can you paste the full error. Normally it is very clear

#

Or it auto fixes the style and can just add the changes and try again

tropic musk Apr 11, 2026, 8:12 PM

#

how can I enable "auto fix"?

pseudo fossil Apr 11, 2026, 8:12 PM

#

It normally does that

tropic musk Apr 11, 2026, 8:13 PM

#

I tried to push to a new branch and it got rejected there with no option to fix anything

pseudo fossil Apr 11, 2026, 8:13 PM

#

Paste the full error message

#

And command

#

Did you fork the repo or just clone it?

tropic musk Apr 11, 2026, 8:14 PM

#

I tried cloning, creating a new branch and that's where I got rejected

#

so I forked it and the code is sitting in my fork

pseudo fossil Apr 11, 2026, 8:14 PM

#

Ok good

#

You can't create a PR from your fork?

tropic musk Apr 11, 2026, 8:15 PM

#

let me try

pseudo fossil Apr 11, 2026, 8:16 PM

#

You are not an ESPHome maintainer so don't have permissions to create a branch in the actual repo. Have to create one in your fork and create a PR there

tropic musk Apr 11, 2026, 8:16 PM

#

I need to reset everything, I have different code in the fork

#

creating a new branch on my fork is giving some formatting errors

#

ruff (legacy alias)......................................................Failed
- hook id: ruff
- files were modified by this hook

Found 2 errors (2 fixed, 0 remaining).

ruff format..............................................................Passed
flake8...................................................................Passed
don't commit to branch...................................................Passed
fix end of files.........................................................Passed
trim trailing whitespace.................................................Failed
- hook id: trailing-whitespace
- exit code: 1
- files were modified by this hook

Fixing esphome/components/bb_epaper/bb_epaperbase.cpp

pyupgrade................................................................Passed
yamllint.............................................(no files to check)Skipped
clang-format.............................................................Failed
- hook id: clang-format
- files were modified by this hook
pylint...................................................................Passed
Update clang-tidy hash...............................(no files to check)Skipped

#

looks like it tried to fix them

pseudo fossil Apr 11, 2026, 8:19 PM

#

That looks good

tropic musk Apr 11, 2026, 8:19 PM

#

branch created

pseudo fossil Apr 11, 2026, 8:19 PM

#

Perfect, can now create a PR

#

https://github.com/esphome/esphome/pulls

GitHub

Pull requests · esphome/esphome

ESPHome is a system to control your ESP32, ESP8266, BK72xx, RP2040 by simple yet powerful configuration files and control them remotely through Home Automation systems. - Pull requests · esphome/es...

#

Will have an option there (if you recently pushed) or under your fork

tropic musk Apr 11, 2026, 8:25 PM

#

the PR was using the wrong branch. Not sure how to make it work now

pseudo fossil Apr 11, 2026, 8:28 PM

#

It looked fine to me

#

Merging into dev from bitbank2:bb_epaper

tropic musk Apr 11, 2026, 8:29 PM

#

no, that PR picked up my bb_temp component

#

let me start from a clean fork

#

https://github.com/esphome/esphome/pull/15652

#

It's complaining that there's no codeowners in init.py, but there is ??

#bb_paper inclusion in ESPHome