Struct PostgresDriver 

pub struct PostgresDriver { /* private fields */ }

Postgres implementation of MigrationDriver.

The pool must have at least 2 connections: one is reserved for the advisory lock, and at least one more is needed for migration queries. PgPoolOptions::new().max_connections(2) is the recommended minimum for CLI use.

Implementations

impl PostgresDriver

pub fn new(pool: PgPool, config: PostgresConfig) -> Result<Self>

Create a new PostgresDriver.

Errors

Returns Error::InvalidIdentifier if the table or schema name contains characters outside [A-Za-z0-9_].

Returns Error::PoolTooSmall if the pool has fewer than 2 connections. One connection is reserved for the advisory lock; all migration queries need at least one more. With max_connections == 1 every pool.begin() call would wait forever.

Trait Implementations

impl Debug for PostgresDriver

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl MigrationDriver for PostgresDriver

fn acquire_lock<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = Result<Box<dyn LockGuard>>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Acquire an advisory lock. The lock is held until the returned guard is dropped.

fn run_setup_sql<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, name: &'life1 str, sql: &'life2 str, ) -> Pin<Box<dyn Future<Output = Result<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait,

Execute one 00_setup file’s SQL in a single transaction (untracked).

fn ensure_tracking_table<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = Result<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Ensure the deployment tracking table (and schema, if configured) exist.

fn applied<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = Result<Vec<AppliedMigration>>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Return all applied migrations, ordered by (version ASC, file ASC).

fn apply<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, migration: &'life1 Migration, up_sql: &'life2 str, batch: i32, ) -> Pin<Box<dyn Future<Output = Result<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait,

Apply a migration: run UP SQL AND insert the tracking row in ONE transaction. The up_sql is passed in rather than re-reading the file in the driver.

fn revert<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, applied: &'life1 AppliedMigration, down_sql: &'life2 str, ) -> Pin<Box<dyn Future<Output = Result<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait,

Revert an applied migration: run down_sql AND delete the tracking row in ONE transaction.