It is possible to use auto-generated models with rayon?

[qarmin]

When having such, auto-generated struct

 # [derive (Default , PartialEq , Debug , Clone)] pub struct r#MainListModel {
         pub r#checked : bool , pub r#filled_header_row : bool , pub r#header_row : bool , pub r#selected_row : bool , pub r#val_int : sp :: ModelRc < i32 > , pub r#val_str : sp :: ModelRc < sp :: SharedString > }
     

and this code

use rayon::prelude::*;
        let main_list_model = MainListModel {
            checked: false,
            filled_header_row: false,
            header_row: false,
            selected_row: false,
            val_int: Default::default(),
            val_str: Default::default(),
        };
        let vc = vec![main_list_model; 10];
        let _ = vc.into_iter().map(|e| e.checked).collect::<Vec<_>>(); // No-problem
        let _ = vc.into_par_iter().map(|e| e.filled_header_row).collect::<Vec<_>>(); // Cannot compile

I have an error

error[E0599]: the method `into_par_iter` exists for struct `Vec<MainListModel>`, but its trait bounds were not satisfied
  --> krokiet/src/connect_delete.rs:95:20
   |
95 |         let _ = vc.into_par_iter().map(|e| e.filled_header_row).collect::<Vec<_>>(); // Cannot compile
   |                    ^^^^^^^^^^^^^
   |
   = note: the following trait bounds were not satisfied:
           `[slint_generatedMainWindow::MainListModel]: Sized`
           which is required by `[slint_generatedMainWindow::MainListModel]: rayon::iter::IntoParallelIterator`
           `[slint_generatedMainWindow::MainListModel]: rayon::iter::ParallelIterator`
           which is required by `[slint_generatedMainWindow::MainListModel]: rayon::iter::IntoParallelIterator`

For more information about this error, try `rustc --explain E0599`.
error: could not compile `krokiet` (bin "krokiet") due to 1 previous error

I think that problematic is, that model definition is

pub struct ModelRc<T>(Option<Rc<dyn Model<Data = T>>>);

which is not Sized

as a workaround, I extract the necessary information from the struct into a separate Vec, use it in parallel, then sort the results to match the original order and remove the filtered-out entries. However, this approach is a bit inefficient

[ogoffart]

I think It is Sized. The problem is most likely that it is not Send
And indeed, we can't send the model to another thread because the notification mechanism can't work across threads.