Parse vector where each entry is parametered by previous vector

[vizigr0u]

Hello,
I'm trying to parse a vector of different sizes data where the type of data is defined in a previous vector of headers, like so :

#[binread]
struct Data {
    #[br(temp)]
    _headers_count: u32,
    #[br(count = _headers_count)]
    pub headers: Vec<MyObjectHeader>,
    #[br(temp, assert(_object_count == _headers_count))]
    _object_count: u32,
    #[br(count = _object_count)]
    pub objects: Vec<MyObject>,
}

#[derive(BinRead)]
struct MyObjectHeader {
    pub object_type: u32,
    // other header data
}

#[derive(BinRead)]
#[br(import(is_a: bool))]
struct MyObject {
    #[br(if(is_a))]
    pub a: Option<A>,
    #[br(if(!is_a))]
    pub b: Option<B>,
}


#[derive(BinRead)]
struct A {
    pub blah: u32,
}

#[derive(BinRead)]
struct B {
    pub foo: u32,
    pub bar: u32,
}

I can't figure out how to pass the data for each element to the parser that parses MyObject, or even the whole header vector and index would work

[csnover]

Hi, thanks for your question!

You are looking for the args_iter helper, which will be used something like this:

#[binread]
struct Data {
    #[br(temp)]
    _headers_count: u32,
    #[br(count = _headers_count)]
    pub headers: Vec<MyObjectHeader>,
    #[br(temp, assert(_object_count == _headers_count))]
    _object_count: u32,
    #[br(parse_with = binrw::helpers::args_iter(&headers))]
    pub objects: Vec<MyObject>,
}

// …

#[derive(BinRead)]
#[br(import_raw(header: &MyObjectHeader))]
struct MyObject {
    // …
}

Due to a bug in the current version of binrw, you will have to do one of two things instead temporarily:

1. Use args_iter_with(&headers, |r, e, a| MyObject::read_options(r, e, a))) instead of args_iter(&headers); or
2. Use args_iter(headers.iter().cloned()) and import_raw(header: MyObjectHeader).

This will be fixed in the next release, which should come out fairly soon, at which point you will be able to do as shown in the original suggestion.