rustyroad 1.8.6

Rusty Road is a framework written in Rust that is based on Ruby on Rails. It is designed to provide the familiar conventions and ease of use of Ruby on Rails, while also taking advantage of the performance and efficiency of Rust.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33

//! Reusable audit-column conventions, not application table/DTO definitions.

use crate::database::introspection::{Column, Table};

/// Selects the transaction-clock column; e.g. a timestamp `updated_at` is eligible.
pub(super) fn clock(table: &Table) -> Option<&Column> {
    table.column("updated_at").filter(|column| {
        timestamp(column) && !column.auto_increment
            && !table.primary_key.contains(&column.name)
    })
}

/// Identifies timestamp types; e.g. `timestamp(3) with time zone` is temporal.
fn timestamp(column: &Column) -> bool {
    column.sql_type.trim().starts_with("timestamp")
}

/// Snapshot payload; e.g. a key plus URL/JSONB/audit columns yields only URL/JSONB.
/// Never replace generated identities, primary/conflict keys, or audit timestamps.
pub(super) fn payload<'a>(table: &'a Table, key: &[String]) -> Vec<&'a Column> {
    table.columns.iter().filter(|column| {
        !column.auto_increment && !table.primary_key.contains(&column.name)
            && !key.contains(&column.name)
            && column.name != "created_at" && column.name != "updated_at"
            && !(timestamp(column) && column.default.is_some())
    }).collect()
}

/// Supports latest-by-text reads; e.g. a nullable `page_url: text` remains eligible.
pub(super) fn text(column: &Column) -> bool {
    let kind = column.sql_type.trim();
    kind == "text" || kind.starts_with("character varying") || kind.starts_with("varchar")
}