Pointing device #9

Univa · 2024-04-04T23:31:28Z

Applicable Changes

Feature addition
Modification to behaviour of existing functionality (potential breaking change)
Modification to existing public APIs (potential breaking change)
Bug fix
User docs modification
Rust toolchain updates
Dependency updates
Workflow change (GitHub Actions)
Repository housekeeping (e.g. READMEs, licensing, templates, etc.)

Related Issues

Description of Changes

Added a pointing module to support pointing devices like a mouse or touchpad. Note that this only adds support for pointing devices over USB for now. Bluetooth support to come later.

Testing Instructions

Try out the IQS5xx pointing driver in your main.rs (assuming a non-split keyboard):

#[keyboard(
    usb,
    pointer(
        driver_setup_fn = test_pointer,
    )
)]
struct MyKeyboard;

impl rumcake::pointer::PointingDevice for MyKeyboard {
    type MouseEventCollector = Self;
}
impl rumcake::pointer::MouseEventCollector for MyKeyboard {}
async fn test_pointer() -> impl rumcake::pointer::PointingDriver {
    rumcake::drivers::iqs5xx::setup_driver(
        setup_i2c! {
            /* ... */
        },
        input_pin!( /* pin identifier */ ),
        output_pin!( /* pin identifier */ ),
    )
}

The touchpad should be capable of basic cursor movements, scrolling (with two fingers sliding), left-clicks (via single finger tap), right-clicks (via two finger tap), and left-click holding (press hold).

More complicated behaviour can be implemented by overriding the implementation of tick and handle_event in the IQS5xxPointingDriver trait

Check-list

I have updated or added to the user docs where applicable
I have tested the changes on real hardware if applicable
The changes do not break compilations with other potentially related feature flags
Credit given where applicable

Univa · 2024-04-04T23:47:34Z

@wallamide Would you be able to test the IQS5xx driver? I don't have access to a touchpad, so whatever I added in here is just best effort. I didn't aim to replicate the implementation that you provided for now, since I'm not sure whether that would be a sane default behaviour. If you would like to have functionality similar to that, you can always reimplement IQS5xxPointingDriver yourself with your own types (particularly handle_event in the trait allows you to register touchpad events however you want).

…`setup_i2c` macro

wallamide · 2024-04-05T17:44:09Z

@Univa thanks for the ping! i'm sadly recovering from being sick so i'll need to check this out in a day or so.

looking this over, i had a few questions:

it looks like i'll need to connect rdy/scl/sda/3.3v/gnd/rst on the trackpad to my mcu, correct?
this should be compatible with a nice!nano v2 build right?
is there via(l) compatibility with the pointer?

thanks again for sharing and i'm looking forward to checking this out!

Univa · 2024-04-05T18:56:16Z

@Univa thanks for the ping! i'm sadly recovering from being sick so i'll need to check this out in a day or so.

looking this over, i had a few questions:
* it looks like i'll need to connect rdy/scl/sda/3.3v/gnd/rst on the trackpad to my mcu, correct?

* this should be compatible with a nice!nano v2 build right?

* is there via(l) compatibility with the pointer?
thanks again for sharing and i'm looking forward to checking this out!

Yes, then you can use the setup_i2c! macro, input_pin! for the RDY pin and output_pin! for the RST pin (shown in the instructions above)
Theoretically, yes. Arguments for setup_i2c on an nRF chip can be found here.
No built-in Via(l) compatibility for now. If you're using Via (not Vial), you can probably implement it in your own workspace by implementing handle_custom_value_command in the ViaKeyboard trait, and adding a custom UI in your definition.json. Docs here: https://univa.github.io/rumcake/api/nrf52840/rumcake/via/trait.ViaKeyboard.html#method.handle_custom_value_command

wallamide · 2024-04-06T22:21:18Z

@Univa thanks again for answering my questions! one last question before i get to this: is there anything preventing me from using the trackpad with a split keyboard on the peripheral/right half? like on zmk for a while you could only add a pointer to the main half not to the peripheral.

just to let you know my plan, i'll be testing out the default functionality with a firmware of just the pointer and giving feedback then i was hoping to try flashing a my layout and making sure it's still working the same. then i'd like to get familiar with the default for a bit and try modifying it.

Univa · 2024-04-07T02:53:40Z

@Univa thanks again for answering my questions! one last question before i get to this: is there anything preventing me from using the trackpad with a split keyboard on the peripheral/right half? like on zmk for a while you could only add a pointer to the main half not to the peripheral.

just to let you know my plan, i'll be testing out the default functionality with a firmware of just the pointer and giving feedback then i was hoping to try flashing a my layout and making sure it's still working the same. then i'd like to get familiar with the default for a bit and try modifying it.

Usage of the trackpad isn't supported on a peripheral side yet since there is no enum variant in MessageToCentral that would be responsible for carrying the input data. Adding support for that shouldn't be too difficult, though. I can get around to that soon.

wallamide · 2024-04-08T08:05:08Z

i was unable to get my firmware to build. i used cargo generate to create a test firmware and i keep getting this error when trying to build:

error: failed to run custom build command for `rust-lzma v0.6.0`

Caused by:
  process didn't exit successfully: `C:\...\pointer-test\target\debug\build\rust-lzma-61258936f45e49e9\build-script-build` (exit code: 101)
  --- stderr
  thread 'main' panicked at C:\Users\...\.cargo\registry\src\index.crates.io-6f17d22bba15001f\rust-lzma-0.6.0\build.rs:13:14:
  Could not find liblzma using vcpkg: VcpkgNotFound("No vcpkg installation found. Set the VCPKG_ROOT environment variable or run 'vcpkg integrate install'")
  note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

i'm running on windows 11.

Univa · 2024-04-08T13:30:47Z

I don't develop on windows, but it looks like you might be missing vcpkg?

also does some slight api changes: - CentralDevice::Layout is now optional - MATRIX_EVENTS get published events from the matrix polling task and central task, instead of peripheral task and layout collect task - split setup function for iqs5xx into two different types: one for default behavior, and another for custom implementor of IQS5xxEventHandler - mouse tasks are now split into two: polling task and mouse event collection task (similar to matrix polling and layout collect) - PeripheralDevice::get_matrix_events_channel changed to return a channel of MessageToCentral instead

Univa · 2024-04-10T20:09:11Z

@wallamide I added support for a peripheral-side pointing device, but it may not be entirely stable. The last commit also changes some of the API, so I also updated the instructions in the original comment.

Peripheral side should look something like this:

#[keyboard(
    usb,
    split_peripheral(
        driver_setup_fn = setup_split_driver
    ),
    pointer(
        driver_setup_fn = test_pointer,
    )
)]
struct MyKeyboard;

async fn test_pointer() -> impl rumcake::pointer::PointingDriver {
    // ...
}
impl rumcake::pointer::PointingDevice for MyKeyboard {
    type PeripheralDeviceType = Self; // send mouse events to a central device
}

impl rumcake::split::peripheral::PeripheralDevice for MyKeyboard {}
async fn setup_split_driver() -> impl rumcake::split::peripheral::PeripheralDeviceDriver {
    // ...
}

Central should look like:

#[keyboard(
    usb,
    split_central(
        driver_setup_fn = setup_split_driver
    )
)]
struct MyKeyboard;

// collects mouse events from peripherals
impl rumcake::pointer::MouseEventCollector for MyKeyboard {}

impl rumcake::split::central::CentralDevice for MyKeyboard {
    type MouseEventCollector = Self;
}
async fn setup_split_driver() -> impl rumcake::split::central::CentralDeviceDriver {
    // ...
}

wallamide · 2024-04-11T01:40:59Z

thanks for the update! i actually am unable to build a basic firmware using the template for the split, though it's looking like this is the result of the refactor. i'll see if i can get it compiling using just the example code in a sec. were you able to check my repo linked above to make sure what i had originally should be working?

Univa · 2024-04-11T13:35:41Z

My bad, I totally forgot to update those templates. I pushed some updates which should hopefully work now. If you use those templates, then add my example code from my previous comment, you should be able to get something compiling.

wallamide · 2024-04-14T19:20:34Z

okay, so i'm unsure what i need to change atm, but i'm getting a few errors from this pointer setup function:

async fn test_pointer() -> impl rumcake::pointer::PointingDriver {
    rumcake::drivers::iqs5xx::setup_driver(
        setup_i2c! {
            interrupt: P0_22,
            i2c: P0_24,
            sda: P0_17,
            scl: P0_20,
        }, 
        input_pin!(P0_24),
        output_pin!(P0_22),
    )
}

here's the main errors, i can post the full log if you need as well. i definitely recommend adding an i2c docs page when you have time or i could help once we can get this sorted out.

error[E0412]: cannot find type `P0_22` in module `$crate::interrupt::typelevel`
error[E0425]: cannot find value `P0_22` in module `rumcake::hw::platform::embassy_nrf::interrupt`
error[E0277]: the trait bound `P0_24: rumcake::hw::platform::embassy_nrf::twim::Instance` is not satisfied

Univa · 2024-04-14T19:41:04Z

i2c should be one of the TWISPI peripherals: https://docs.embassy.dev/embassy-nrf/git/nrf52840/twim/trait.Instance.html#implementors

interrupt should be the corresponding interrupts for that chosen i2c peripheral: https://docs.embassy.dev/embassy-nrf/git/nrf52840/twim/trait.Instance.html#implementors

So, you can change the setup_i2c to something like this:

setup_i2c! {
    interrupt: SPIM0_SPIS0_TWIM0_TWIS0_SPI0_TWI0,
    i2c: TWISPI0,
    sda: P0_17,
    scl: P0_20,
},

I'll add docs to the API reference at some point to hopefully clarify this stuff.

Univa added 5 commits April 4, 2024 19:04

pointing device support + iqs5xx

a2f4a49

add some doc comments for the iqs5xx driver

a200ff0

make iqs5xx driver trait pub

6514347

rename trait to be consistent with pointer module

bb0411d

make iqs5xx driver pub

1dfe41d

make it possible to create blocking i2c implementation for the stm32 …

3749adc

…`setup_i2c` macro

Univa linked an issue Apr 4, 2024 that may be closed by this pull request

add touchpad support #8

Open

4 tasks

Univa added 2 commits April 7, 2024 10:48

Merge branch 'main' into pointing-device

a4b8ee4

missing complement

d47f008

Univa force-pushed the pointing-device branch from 842a2d4 to b6707de Compare April 11, 2024 14:00

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Pointing device #9

Pointing device #9

Univa commented Apr 4, 2024 •

edited

Loading

Univa commented Apr 4, 2024

wallamide commented Apr 5, 2024

Univa commented Apr 5, 2024 •

edited

Loading

wallamide commented Apr 6, 2024 •

edited

Loading

Univa commented Apr 7, 2024

wallamide commented Apr 8, 2024

Univa commented Apr 8, 2024

Univa commented Apr 10, 2024 •

edited

Loading

wallamide commented Apr 11, 2024

Univa commented Apr 11, 2024 •

edited

Loading

wallamide commented Apr 14, 2024

Univa commented Apr 14, 2024

Pointing device #9

Are you sure you want to change the base?

Pointing device #9

Conversation

Univa commented Apr 4, 2024 • edited Loading

Applicable Changes

Related Issues

Description of Changes

Testing Instructions

Check-list

Univa commented Apr 4, 2024

wallamide commented Apr 5, 2024

Univa commented Apr 5, 2024 • edited Loading

wallamide commented Apr 6, 2024 • edited Loading

Univa commented Apr 7, 2024

wallamide commented Apr 8, 2024

Univa commented Apr 8, 2024

Univa commented Apr 10, 2024 • edited Loading

wallamide commented Apr 11, 2024

Univa commented Apr 11, 2024 • edited Loading

wallamide commented Apr 14, 2024

Univa commented Apr 14, 2024

Univa commented Apr 4, 2024 •

edited

Loading

Univa commented Apr 5, 2024 •

edited

Loading

wallamide commented Apr 6, 2024 •

edited

Loading

Univa commented Apr 10, 2024 •

edited

Loading

Univa commented Apr 11, 2024 •

edited

Loading