diesel::query_builder::Query实现不够通用编译错误求助

问题背景 #

在使用Rocket框架结合Diesel ORM构建分页查询时，当用into_boxed()构建动态条件查询后，调用自定义的paginate()和load_and_count_total()方法会触发编译错误，提示diesel::query_builder::Query实现不够通用，具体错误指向BoxedSelectStatement的生命周期约束不满足宏的通用要求。

相关代码 #

Rocket处理器函数 #

#[get("/")]
async fn index(mut db: Db) -> String {
    let query = schema::posts::table.into_boxed();

    let (items, total): (Vec<models::Post>, i64) = query
        .paginate(Some(0))
        .load_and_count_total(&mut db)
        .await
        .unwrap();

    format!("{items:?} {total}")
}

自定义分页Trait及实现 #

pub trait Paginate: Sized {
    fn paginate(self, page: Option<i64>) -> Paginated<Self>;
}

impl<T: Query> Paginate for T {
    fn paginate(self, page: Option<i64>) -> Paginated<Self> {
        let page = page.unwrap_or(0);

        Paginated {
            query: self,
            per_page: DEFAULT_PER_PAGE,
            page,
            offset: page * DEFAULT_PER_PAGE,
        }
    }
}

impl<T> Paginated<T> {
    pub async fn load_and_count_total<'a, U>(
        self,
        conn: &mut AsyncPgConnection,
    ) -> QueryResult<(Vec<U>, i64)>
    where
        Self: LoadQuery<'a, AsyncPgConnection, (U, i64)>,
        U: Send,
        T: 'a,
    {
        #![allow(clippy::get_first)]

        let results = self.load::<(U, i64)>(conn).await?;
        let total = results.get(0).map(|x| x.1).unwrap_or(0);
        let records = results.into_iter().map(|x| x.0).collect();
        Ok((records, total))
    }
}

编译错误信息 #

error: implementation of diesel::query_builder::Query is not general enough
--> src/main.rs:12:1
|

#[get("/")]
^^^^^^^^^^^ implementation of diesel::query_builder::Query is not general enough

= note: diesel::query_builder::Query would have to be implemented for the type BoxedSelectStatement<'0, (Integer, Text), FromClause<table>, Pg>, for any lifetime '0...
= note: ...but diesel::query_builder::Query is actually implemented for the type BoxedSelectStatement<'1, (Integer, Text), FromClause<table>, Pg>, for some specific lifetime '1
= note: this error originates in the attribute macro get (in Nightly builds, run with -Z macro-backtrace for more info)

解决方案 #

问题核心是load_and_count_total的生命周期约束过于严格，仅要求Self对特定生命周期'a实现LoadQuery，但Rocket路由宏需要类型满足任意生命周期的通用性要求。通过**更高阶生命周期（HRTB）**调整约束即可解决。

修改Paginated的load_and_count_total方法：

impl<T> Paginated<T> {
    pub async fn load_and_count_total<U>(
        self,
        conn: &mut AsyncPgConnection,
    ) -> QueryResult<(Vec<U>, i64)>
    where
        // 使用HRTB要求Self对任意连接生命周期'conn都实现LoadQuery
        for<'conn> Self: LoadQuery<'conn, AsyncPgConnection, (U, i64)>,
        U: Send,
    {
        #![allow(clippy::get_first)]

        let results = self.load::<(U, i64)>(conn).await?;
        let total = results.get(0).map(|x| x.1).unwrap_or(0);
        let records = results.into_iter().map(|x| x.0).collect();
        Ok((records, total))
    }
}

关键修改点 #

1. 移除方法上的'a生命周期参数和T: 'a约束
2. 将Self: LoadQuery<'a, ...>替换为for<'conn> Self: LoadQuery<'conn, ...>，明确要求Self对任意连接生命周期都能实现LoadQuery，满足Rocket宏对类型通用性的要求

调整后，编译器可正确推导BoxedSelectStatement的生命周期，既能保留into_boxed()的动态查询能力，又能正常使用分页方法。

内容的提问来源于stack exchange，提问作者Hans