On Tue, Dec 10, 2024 at 4:58 PM Zhao Liu <[email protected]> wrote:
> Great idea. It nicely balances the differences between Rust and C QOM
> conventions.
Except it does not work. :( Suppose you have
pub struct MySuperclass {
parent: DeviceState,
field: Box<MyData>,
...
}
impl Drop for MySuperclass {
...
}
pub struct MySubclass {
parent: MySuperclass,
...
}
When instance_finalize is called for MySubclass, it will walk the
struct's list of fields and call the drop method for MySuperclass.
Then, object_deinit recurses to the superclass and calls the same drop
method again. This will cause double-freeing of the Box<MyData>, or
more in general double-dropping.
What's happening here is that QOM wants to control the drop order of
MySuperclass and MySubclass's fields. To do so, the parent field must
be marked ManuallyDrop<>, which is quite ugly. Perhaps we can add a
wrapper type ParentField<> that is specific to QOM. This hides the
implementation detail of *what* is special about the ParentField, and
it will also be easy to check for in the #[derive(Object)] macro.
Maybe in the future it will even make sense to have special functions
implemented on ParentField, I don't know...
Paolo
