embedded-hal v1.0.0 and embedded-hal-bus about ssd1306 HOT 4 OPEN

Could you please try with embedded-hal-bus = "0.1.0" instead of trying to use the dependency from github?

from ssd1306.

Changing Cargo.toml so I have

embedded-hal-bus = "0.1.0"

and text_i2c.rs to what I think I should have:

...
use hal::{
    //i2c::I2c,
    prelude::*,
    pac::{Peripherals, I2C1},
    pac,
};

use embedded_hal_bus::i2c::I2c;
...

Then when I compile I get

...

 Downloaded embedded-hal-bus v0.1.0
  Downloaded 1 crate (11.9 KB) in 0.72s
   Compiling embedded-hal-bus v0.1.0
   Compiling ssd1306 v0.8.4 (/home/paul/githubClones/ssd1306)

error[E0432]: unresolved import `embedded_hal_bus::i2c::I2c`
  --> examples/text_i2c.rs:53:5
   |
53 | use embedded_hal_bus::i2c::I2c;
   |     ^^^^^^^^^^^^^^^^^^^^^^^^^^ no `I2c` in `i2c`
   |
help: consider importing one of these items instead
   |
53 | use crate::hal::i2c::I2c;
   |     ~~~~~~~~~~~~~~~~~~~~
53 | use embedded_hal::i2c::I2c;
   |     ~~~~~~~~~~~~~~~~~~~~~~
53 | use embedded_hal_async::i2c::I2c;
   |     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
53 | use stm32f1xx_hal::i2c::I2c;
   |     ~~~~~~~~~~~~~~~~~~~~~~~
     and 2 other candidates

For more information about this error, try `rustc --explain E0432`.
error: could not compile `ssd1306` (example "text_i2c") due to 1 previous error

I find it strange that there is "no I2c in i2c". It really looks to me like it should be there. I don't see typos or anything simple, but this does not seem right.

Also, when I look at cargo tree, it looks like display-interface-i2c depends on embedded-hal-async but not on embedded-hal-bus.

$ cargo tree
ssd1306 v0.8.4 (/home/paul/githubClones/ssd1306)
├── display-interface v0.5.0 (https://github.com/bugadani/display-interface.git?branch=eh1#e65eaf6d)
├── display-interface-i2c v0.5.0 (https://github.com/bugadani/display-interface.git?branch=eh1#e65eaf6d)
│ ├── display-interface v0.5.0 (https://github.com/bugadani/display-interface.git?branch=eh1#e65eaf6d)
│ ├── embedded-hal v1.0.0
│ └── embedded-hal-async v1.0.0
│ └── embedded-hal v1.0.0
├── display-interface-spi v0.5.0 (https://github.com/bugadani/display-interface.git?branch=eh1#e65eaf6d)
│ ├── byte-slice-cast v1.2.2
│ ├── display-interface v0.5.0 (https://github.com/bugadani/display-interface.git?branch=eh1#e65eaf6d)
│ ├── embedded-hal v1.0.0
│ └── embedded-hal-async v1.0.0 ()
├── embedded-graphics-core v0.4.0
│ ├── az v1.2.1
│ └── byteorder v1.5.0
├── embedded-hal v1.0.0
├── embedded-hal-async v1.0.0 ()
└── embedded-hal-bus v0.1.0
├── critical-section v1.1.2
└── embedded-hal v1.0.0
[dev-dependencies]
...

from ssd1306.

embedded-hal-bus isn't a simple plug-and-play crate, you don't get an I2c struct that "just works". You need to decide on how you share your bus between your devices, and for that you have three options.

Why do you expect display-interface-i2c to depend on embedded-hal-bus? embedded-hal-bus doesn't define any new traits that would require any special handling in display-interface-i2c.

from ssd1306.

Thanks for the hint @bugadani . I guess it is more complicated than I thought. I'll eventually figure it out. Not exactly a simple replacement for shared_bus, but seems like more control over the mechanism.

My thinking on display-interface-i2c depending on embedded-hal-bus was because I mis-read the tree and thought it depended on embedded-hal-spi. I think the spi and i2c similarities are such that one dependency would suggest the other too..

from ssd1306.