rdbms-playground/tests/it/sql_dml_e2e.rs

//! Sub-phase 3k — Tier-3 end-to-end DML integration tests
//! (ADR-0033, plan `docs/plans/20260520-adr-0033-phase-3.md`
//! "Sub-phase 3k").
//!
//! Where the per-sub-phase `tests/sql_{insert,update,delete}.rs`
//! suites drive the worker directly with hand-written arguments,
//! these tests exercise the **full advanced-mode path**: a literal
//! line is parsed in Advanced mode (the same `parse_command`
//! dispatch the runtime uses), the resulting `Command::Sql*` is
//! executed through the worker, and the persisted CSV / history /
//! result set are asserted. They cover the real-world DML shapes
//! the 3k exit gate lists:
//!
//! - `INSERT … SELECT` cross-table
//! - multi-row `INSERT` covering all ten playground types, with
//!   `RETURNING` recovering every type (matrix R5)
//! - `UPDATE` with a subquery in `SET`
//! - `DELETE` with cascade (per-relationship summary + multi-table
//!   re-persistence)
//! - `UPSERT` round-trip (`DO UPDATE` then `DO NOTHING`)
//! - `RETURNING` on each of `INSERT` / `UPDATE` / `DELETE`
//! - `history.log` replay of every Phase-3 statement form
//! - the OOS parse-rejections (ADR-0033 §13)
//! - the `[ERR]`/`[WRN]` validity indicator firing on a SQL DML
//!   diagnostic (matrix A7)

use ratatui::Terminal;
use ratatui::backend::TestBackend;

use rdbms_playground::app::App;
use rdbms_playground::db::{Database, DbError, DeleteResult, InsertResult, UpdateResult};
use rdbms_playground::dsl::parser::parse_command_in_mode;
use rdbms_playground::dsl::walker::Severity;
use rdbms_playground::dsl::{
    ColumnSpec, Command, ReferentialAction, RowFilter, Type, parse_command,
};
use rdbms_playground::event::AppEvent;
use rdbms_playground::mode::Mode;
use rdbms_playground::persistence::Persistence;
use rdbms_playground::project;
use rdbms_playground::runtime::run_replay;
use rdbms_playground::theme::Theme;
use rdbms_playground::ui;

// ---------------------------------------------------------------
// Harness — mirrors the per-sub-phase suites' helpers.
// ---------------------------------------------------------------

fn rt() -> tokio::runtime::Runtime {
    tokio::runtime::Builder::new_current_thread()
        .enable_all()
        .build()
        .expect("tokio rt")
}

fn open_project_db() -> (project::Project, Database, tempfile::TempDir) {
    let dir = tempfile::tempdir().expect("create tempdir");
    let project =
        project::open_or_create(None, Some(dir.path())).expect("open or create project");
    let persistence = Persistence::new(project.path().to_path_buf());
    let db = Database::open_with_persistence(project.db_path(), persistence)
        .expect("open db with persistence");
    (project, db, dir)
}

fn read_csv(project: &project::Project, table: &str) -> Option<String> {
    std::fs::read_to_string(project.path().join("data").join(format!("{table}.csv"))).ok()
}

fn create_cols(
    db: &Database,
    rt: &tokio::runtime::Runtime,
    name: &str,
    cols: &[(&str, Type)],
    pk: &[&str],
) {
    rt.block_on(db.create_table(
        name.to_string(),
        cols.iter().map(|(n, t)| ColumnSpec::new(*n, *t)).collect(),
        pk.iter().map(|s| (*s).to_string()).collect(),
        None,
    ))
    .unwrap_or_else(|e| panic!("create table {name}: {e:?}"));
}

/// Parse `input` in Advanced mode and run the resulting SQL INSERT
/// through the worker — the full parse → execute path.
fn run_insert(
    db: &Database,
    rt: &tokio::runtime::Runtime,
    input: &str,
) -> Result<InsertResult, DbError> {
    match parse_command(input).unwrap_or_else(|e| panic!("parse {input:?}: {e:?}")) {
        Command::SqlInsert {
            sql,
            target_table,
            listed_columns,
            row_source,
            returning,
            ..
        } => rt.block_on(db.run_sql_insert(
            sql,
            Some(input.to_string()),
            target_table,
            listed_columns,
            row_source,
            returning,
        )),
        other => panic!("expected Command::SqlInsert from {input:?}, got {other:?}"),
    }
}

fn run_update(
    db: &Database,
    rt: &tokio::runtime::Runtime,
    input: &str,
) -> Result<UpdateResult, DbError> {
    match parse_command(input).unwrap_or_else(|e| panic!("parse {input:?}: {e:?}")) {
        Command::SqlUpdate { sql, target_table, returning, set_literals } => rt.block_on(
            db.run_sql_update_with_literals(
                sql,
                Some(input.to_string()),
                target_table,
                returning,
                set_literals,
            ),
        ),
        other => panic!("expected Command::SqlUpdate from {input:?}, got {other:?}"),
    }
}

fn run_delete(
    db: &Database,
    rt: &tokio::runtime::Runtime,
    input: &str,
) -> Result<DeleteResult, DbError> {
    match parse_command(input).unwrap_or_else(|e| panic!("parse {input:?}: {e:?}")) {
        Command::SqlDelete { sql, target_table, returning } => rt.block_on(
            db.run_sql_delete(sql, Some(input.to_string()), target_table, returning),
        ),
        other => panic!("expected Command::SqlDelete from {input:?}, got {other:?}"),
    }
}

/// Seed rows through the SQL INSERT path (no auto-gen columns, so
/// the statement executes verbatim).
fn seed(db: &Database, rt: &tokio::runtime::Runtime, sql: &str) {
    run_insert(db, rt, sql).unwrap_or_else(|e| panic!("seed {sql:?}: {e:?}"));
}

fn query(db: &Database, rt: &tokio::runtime::Runtime, table: &str) -> Vec<Vec<Option<String>>> {
    rt.block_on(db.query_data(table.to_string(), None, None, None))
        .unwrap_or_else(|e| panic!("query_data {table}: {e:?}"))
        .rows
}

// ===============================================================
// INSERT … SELECT cross-table
// ===============================================================

#[test]
fn e2e_insert_select_cross_table_copies_rows_and_persists_both() {
    let (project, db, _dir) = open_project_db();
    let rt = rt();
    create_cols(&db, &rt, "source", &[("id", Type::Int), ("v", Type::Text)], &["id"]);
    create_cols(&db, &rt, "archive", &[("id", Type::Int), ("v", Type::Text)], &["id"]);
    seed(&db, &rt, "insert into source (id, v) values (1, 'x'), (2, 'y')");

    let result = run_insert(&db, &rt, "insert into archive select * from source")
        .expect("INSERT … SELECT runs");
    assert_eq!(result.rows_affected, 2, "two source rows copied");

    let archive_csv = read_csv(&project, "archive").expect("archive.csv");
    assert!(
        archive_csv.contains('x') && archive_csv.contains('y'),
        "archive reflects both copied rows: {archive_csv:?}",
    );
    let source_csv = read_csv(&project, "source").expect("source.csv");
    assert!(
        source_csv.contains('x') && source_csv.contains('y'),
        "source is left intact: {source_csv:?}",
    );
}

// ===============================================================
// Multi-row INSERT covering all ten playground types + RETURNING
// type recovery for every type (matrix R5).
// ===============================================================

#[test]
fn e2e_multirow_insert_all_ten_types_roundtrips_and_returning_recovers_each_type() {
    let (project, db, _dir) = open_project_db();
    let rt = rt();
    // serial PK + shortid auto-fill; the other eight columns are
    // user-supplied. `blob` has no value-literal grammar yet
    // (see src/dsl/value.rs), so it is inserted NULL — its *type*
    // still round-trips through the RETURNING column-origin path.
    create_cols(
        &db,
        &rt,
        "allten",
        &[
            ("ser", Type::Serial),
            ("txt", Type::Text),
            ("i", Type::Int),
            ("r", Type::Real),
            ("dec", Type::Decimal),
            ("flag", Type::Bool),
            ("d", Type::Date),
            ("ts", Type::DateTime),
            ("bl", Type::Blob),
            ("sid", Type::ShortId),
        ],
        &["ser"],
    );

    let result = run_insert(
        &db,
        &rt,
        "insert into allten (txt, i, r, dec, flag, d, ts, bl) values \
         ('hi', 42, 1.5, 9.50, true, '2026-05-23', '2026-05-23 10:00:00', null), \
         ('yo', 7, 2.5, 3.25, false, '2025-01-01', '2025-01-01 00:00:00', null) \
         returning ser, txt, i, r, dec, flag, d, ts, bl, sid",
    )
    .expect("multi-row INSERT … RETURNING runs");

    assert_eq!(result.rows_affected, 2, "two rows inserted");
    assert_eq!(result.data.rows.len(), 2, "RETURNING yields both rows");

    // Every one of the ten playground types is recovered via the
    // RETURNING column-origin path (matrix R5).
    assert_eq!(
        result.data.column_types,
        vec![
            Some(Type::Serial),
            Some(Type::Text),
            Some(Type::Int),
            Some(Type::Real),
            Some(Type::Decimal),
            Some(Type::Bool),
            Some(Type::Date),
            Some(Type::DateTime),
            Some(Type::Blob),
            Some(Type::ShortId),
        ],
        "RETURNING recovers each of the ten playground types; got {:?}",
        result.data.column_types,
    );

    // Values round-trip: serial auto-incremented (1, 2), shortid
    // auto-filled (non-empty + distinct), the user values persisted.
    let rows = query(&db, &rt, "allten");
    assert_eq!(rows.len(), 2, "both rows persisted");
    let csv = read_csv(&project, "allten").expect("allten.csv");
    assert!(csv.contains("hi") && csv.contains("yo"), "text round-trips: {csv:?}");
    assert!(csv.contains("2026-05-23") && csv.contains("2025-01-01"), "dates round-trip: {csv:?}");

    let sids: Vec<&str> = rows.iter().filter_map(|r| r[9].as_deref()).collect();
    assert_eq!(sids.len(), 2, "both shortids present");
    assert!(sids.iter().all(|s| !s.is_empty()), "shortids non-empty: {sids:?}");
    assert_ne!(sids[0], sids[1], "auto-filled shortids are distinct: {sids:?}");
    let sers: Vec<&str> = rows.iter().filter_map(|r| r[0].as_deref()).collect();
    assert!(sers.contains(&"1") && sers.contains(&"2"), "serial auto-incremented: {sers:?}");
}

// ===============================================================
// UPDATE with a subquery in SET
// ===============================================================

#[test]
fn e2e_update_with_subquery_in_set() {
    let (project, db, _dir) = open_project_db();
    let rt = rt();
    create_cols(
        &db,
        &rt,
        "customers",
        &[("id", Type::Int), ("name", Type::Text), ("last_order", Type::Int)],
        &["id"],
    );
    create_cols(
        &db,
        &rt,
        "orders",
        &[("id", Type::Int), ("cust", Type::Int), ("amount", Type::Int)],
        &["id"],
    );
    seed(&db, &rt, "insert into customers (id, name, last_order) values (1, 'A', 0), (2, 'B', 0)");
    seed(&db, &rt, "insert into orders (id, cust, amount) values (10, 1, 50), (11, 1, 30), (12, 2, 99)");

    let result = run_update(
        &db,
        &rt,
        "update customers set last_order = \
         (select max(amount) from orders where cust = customers.id)",
    )
    .expect("UPDATE with subquery in SET runs");
    assert_eq!(result.rows_affected, 2, "both customers updated");

    let rows = query(&db, &rt, "customers");
    let c1 = rows.iter().find(|r| r[0].as_deref() == Some("1")).expect("customer 1");
    let c2 = rows.iter().find(|r| r[0].as_deref() == Some("2")).expect("customer 2");
    assert_eq!(c1[2].as_deref(), Some("50"), "customer 1 → max(50, 30) = 50");
    assert_eq!(c2[2].as_deref(), Some("99"), "customer 2 → max(99) = 99");

    let csv = read_csv(&project, "customers").expect("customers.csv");
    assert!(csv.contains("50") && csv.contains("99"), "CSV reflects the update: {csv:?}");
}

// ===============================================================
// DELETE with cascade — per-relationship summary + multi-table
// re-persistence.
// ===============================================================

fn cascade_fixture(db: &Database, rt: &tokio::runtime::Runtime) {
    create_cols(db, rt, "Customers", &[("id", Type::Int), ("Name", Type::Text)], &["id"]);
    create_cols(db, rt, "Orders", &[("id", Type::Int), ("CustId", Type::Int)], &["id"]);
    rt.block_on(db.add_relationship(
        Some("places".to_string()),
        "Customers".to_string(),
        vec!["id".to_string()],
        "Orders".to_string(),
        vec!["CustId".to_string()],
        ReferentialAction::Cascade,
        ReferentialAction::NoAction,
        false,
        None,
    ))
    .expect("add cascade relationship");
    seed(db, rt, "insert into Customers (id, Name) values (1, 'Alice'), (2, 'Bob')");
    seed(db, rt, "insert into Orders (id, CustId) values (10, 1), (11, 1), (12, 2)");
}

#[test]
fn e2e_delete_with_cascade_reports_summary_and_repersists_children() {
    let (project, db, _dir) = open_project_db();
    let rt = rt();
    cascade_fixture(&db, &rt);

    let result = run_delete(&db, &rt, "delete from Customers where id = 1")
        .expect("cascading DELETE runs");
    assert_eq!(result.rows_affected, 1, "one parent row deleted");
    assert_eq!(result.cascade.len(), 1, "one cascade relationship affected");
    let effect = &result.cascade[0];
    assert_eq!(effect.child_table, "Orders");
    assert_eq!(effect.rows_changed, 2, "Alice's two orders cascaded");

    let orders_csv = read_csv(&project, "Orders").expect("Orders.csv re-persisted");
    assert!(orders_csv.contains("12"), "Bob's order (12) preserved: {orders_csv:?}");
    assert!(!orders_csv.contains("10"), "Alice's order 10 cascaded away: {orders_csv:?}");
    assert!(!orders_csv.contains("11"), "Alice's order 11 cascaded away: {orders_csv:?}");
}

// ===============================================================
// UPSERT round-trip — DO UPDATE then DO NOTHING.
// ===============================================================

#[test]
fn e2e_upsert_round_trip_do_update_then_do_nothing() {
    let (project, db, _dir) = open_project_db();
    let rt = rt();
    create_cols(&db, &rt, "kv", &[("id", Type::Int), ("name", Type::Text)], &["id"]);
    seed(&db, &rt, "insert into kv (id, name) values (1, 'old')");

    // DO UPDATE on a conflict mutates the existing row.
    let upd = run_insert(
        &db,
        &rt,
        "insert into kv (id, name) values (1, 'new') on conflict (id) do update set name = excluded.name",
    )
    .expect("UPSERT DO UPDATE runs");
    assert_eq!(upd.rows_affected, 1, "DO UPDATE touches the conflicting row");
    let csv = read_csv(&project, "kv").expect("kv.csv");
    assert!(csv.contains("new") && !csv.contains("old"), "row updated to 'new': {csv:?}");

    // DO NOTHING on a conflict is a no-op.
    let nothing = run_insert(
        &db,
        &rt,
        "insert into kv (id, name) values (1, 'ignored') on conflict (id) do nothing",
    )
    .expect("UPSERT DO NOTHING runs");
    assert_eq!(nothing.rows_affected, 0, "DO NOTHING changes no rows");
    let csv = read_csv(&project, "kv").expect("kv.csv");
    assert!(csv.contains("new") && !csv.contains("ignored"), "row unchanged by DO NOTHING: {csv:?}");
}

// ===============================================================
// RETURNING on each of INSERT / UPDATE / DELETE.
// ===============================================================

#[test]
fn e2e_returning_on_insert_update_delete() {
    let (_project, db, _dir) = open_project_db();
    let rt = rt();
    create_cols(&db, &rt, "t", &[("id", Type::Int), ("v", Type::Text)], &["id"]);

    let ins = run_insert(&db, &rt, "insert into t (id, v) values (1, 'a') returning id, v")
        .expect("INSERT … RETURNING runs");
    assert_eq!(ins.data.rows.len(), 1, "INSERT RETURNING yields the inserted row");
    assert_eq!(ins.data.rows[0][1].as_deref(), Some("a"));

    let upd = run_update(&db, &rt, "update t set v = 'b' where id = 1 returning v")
        .expect("UPDATE … RETURNING runs");
    assert_eq!(upd.data.rows.len(), 1, "UPDATE RETURNING yields the modified row");
    assert_eq!(upd.data.rows[0][0].as_deref(), Some("b"));

    let del = run_delete(&db, &rt, "delete from t where id = 1 returning *")
        .expect("DELETE … RETURNING runs");
    assert_eq!(del.data.rows.len(), 1, "DELETE RETURNING yields the pre-delete row");
    assert_eq!(del.data.rows[0][1].as_deref(), Some("b"), "pre-delete value surfaced");
    assert!(query(&db, &rt, "t").is_empty(), "row is gone after the DELETE");
}

// ===============================================================
// history.log replay of every Phase-3 statement form.
// ===============================================================

#[test]
fn e2e_replay_phase3_dml_forms_from_a_script() {
    let dir = tempfile::tempdir().expect("tempdir");
    let project = project::open_or_create(None, Some(dir.path())).expect("project");
    let db = Database::open_with_persistence(
        project.db_path(),
        Persistence::new(project.path().to_path_buf()),
    )
    .expect("db");
    let rt = rt();

    // A script of Phase-3 SQL DML forms (plus the DDL needed to set
    // up). Replay parses each line in Advanced mode (ADR-0033
    // Amendment 3), so the SQL forms route to the SQL worker path.
    std::fs::write(
        project.path().join("phase3.commands"),
        "create table T with pk id(int)\n\
         add column T: v (text)\n\
         insert into T (id, v) values (1, 'a'), (2, 'b'), (3, 'c')\n\
         insert into T select id + 10, v from T where id = 1\n\
         update T set v = 'z' where id = 2\n\
         delete from T where id = 3\n",
    )
    .expect("write script");

    let events = rt.block_on(run_replay(&db, project.path(), "phase3.commands"));
    match events.last().expect("at least one event") {
        AppEvent::ReplayCompleted { count, .. } => {
            assert_eq!(*count, 6, "all six lines replayed; events: {events:?}");
        }
        other => panic!("expected ReplayCompleted, got {other:?} (events: {events:?})"),
    }

    // Faithful application: multi-row insert + INSERT…SELECT +
    // UPDATE + DELETE all landed.
    let rows = query(&db, &rt, "T");
    let mut by_id: Vec<(String, Option<String>)> = rows
        .iter()
        .map(|r| (r[0].clone().unwrap_or_default(), r[1].clone()))
        .collect();
    by_id.sort();
    assert_eq!(
        by_id,
        vec![
            ("1".to_string(), Some("a".to_string())),
            ("11".to_string(), Some("a".to_string())), // INSERT…SELECT id+10
            ("2".to_string(), Some("z".to_string())),  // UPDATE
            // id 3 was DELETEd
        ]
        .into_iter()
        .collect::<std::collections::BTreeMap<_, _>>()
        .into_iter()
        .collect::<Vec<_>>(),
        "replayed DML applied faithfully; got {by_id:?}",
    );
}

// ===============================================================
// OOS parse-rejections (ADR-0033 §13) — behaviour confirmed; pin it.
// ===============================================================

#[test]
fn e2e_out_of_scope_dml_forms_parse_reject() {
    let cases = [
        ("OOS-1 DEFAULT VALUES", "insert into t default values"),
        ("OOS-2 INSERT OR REPLACE", "insert or replace into t values (1)"),
        ("OOS-2 INSERT OR IGNORE", "insert or ignore into t values (1)"),
        ("OOS-3 UPDATE … FROM", "update t set a = b.x from other b where t.id = b.id"),
        ("OOS-4 WITH … UPDATE", "with x as (select 1) update t set a = 1 where id = 1"),
        ("OOS-4 WITH … DELETE", "with x as (select 1) delete from t where id = 1"),
        ("OOS-5 INDEXED BY", "delete from t indexed by idx where id = 1"),
        ("OOS-5 NOT INDEXED", "update t not indexed set a = 1 where id = 1"),
        ("OOS-6 multi-statement (DELETE; DELETE)", "delete from t where id = 1; delete from t where id = 2"),
        ("OOS-6 multi-statement (INSERT; INSERT)", "insert into t values (1); insert into t values (2)"),
    ];
    for (label, src) in cases {
        assert!(
            parse_command_in_mode(src, Mode::Advanced).is_err(),
            "{label}: {src:?} must parse-reject (ADR-0033 §13)",
        );
    }
}

#[test]
fn e2e_single_dml_statement_with_trailing_semicolon_parses() {
    // Guard for the OOS-6 multi-statement rejection above: a *single*
    // statement with a trailing `;` is still valid (ADR-0033 §1 — the
    // optional `;` tail), so the rejection above is genuinely about a
    // second statement, not the semicolon.
    assert!(
        matches!(
            parse_command_in_mode("delete from t where id = 1;", Mode::Advanced),
            Ok(Command::SqlDelete { .. })
        ),
        "a single statement with a trailing semicolon must still parse",
    );
}

#[test]
fn e2e_update_all_rows_in_advanced_falls_back_to_dsl() {
    // ADR-0033 Amendment 4 reverses Amendment 3's counter-example: the
    // SQL `UPDATE`'s `SET` expression must NOT consume the DSL flag
    // `--all-rows`. An adjacent `--` is not two minus operators — the
    // playground has no `--` line comment — so the SQL shape fails and
    // dispatch falls back to the DSL `Update { AllRows }`, mirroring
    // `delete … --all-rows`.
    assert!(
        matches!(
            parse_command_in_mode("update Orders set total = 42 --all-rows", Mode::Advanced),
            Ok(Command::Update { filter: RowFilter::AllRows, .. })
        ),
        "advanced `update … --all-rows` falls back to the DSL Update",
    );
    // Legitimate spaced arithmetic is unaffected — the dashes are not
    // adjacent, so this stays a SQL UPDATE (total = 42 - (-3) = 45).
    assert!(
        matches!(
            parse_command_in_mode("update Orders set total = 42 - -3", Mode::Advanced),
            Ok(Command::SqlUpdate { .. })
        ),
        "spaced `42 - -3` stays a SQL UPDATE",
    );
    // An adjacent `--` before a number is no longer silently accepted as
    // arithmetic; with no `--all-rows` flag to fall back to, it is a
    // parse error (acceptable per Amendment 4 — contrived input, and the
    // playground does not support `--` comments).
    assert!(
        parse_command_in_mode("update Orders set total = 42--3", Mode::Advanced).is_err(),
        "adjacent `42--3` is a parse error (no `--` comment support)",
    );
}

// ===============================================================
// Validity indicator fires on a SQL DML diagnostic (matrix A7).
// ===============================================================

fn rendered_text(app: &mut App, theme: &Theme, width: u16, height: u16) -> String {
    let backend = TestBackend::new(width, height);
    let mut terminal = Terminal::new(backend).expect("terminal");
    terminal.draw(|f| ui::render(app, theme, f)).expect("draw");
    let buffer = terminal.backend().buffer().clone();
    let mut out = String::new();
    for y in 0..buffer.area.height {
        for x in 0..buffer.area.width {
            out.push_str(buffer[(x, y)].symbol());
        }
        out.push('\n');
    }
    out
}

#[test]
fn e2e_validity_indicator_fires_for_sql_dml_diagnostic() {
    // ADR-0027 §4 / ADR-0030 §8 / matrix A7: a SQL DML line whose
    // WHERE carries a predicate warning (`= NULL`) lights up the
    // `[WRN]` indicator in Advanced mode. The verdict is the same
    // computation the runtime stores in `input_indicator`.
    let mut app = App::new();
    app.mode = Mode::Advanced;
    // Populate the schema cache so the diagnostic pass resolves
    // the column.
    app.schema_cache.tables.push("t".to_string());
    app.schema_cache.columns.push("v".to_string());
    app.schema_cache.table_columns.insert(
        "t".to_string(),
        vec![rdbms_playground::completion::TableColumn {
            name: "v".to_string(),
            user_type: Type::Int,
            not_null: false,
            has_default: false,
        }],
    );
    app.input = "update t set v = 1 where v = NULL".to_string();

    assert_eq!(
        app.input_validity_verdict(),
        Some(Severity::Warning),
        "a SQL DML `= NULL` predicate raises a WARNING verdict",
    );

    // And the indicator renders the `[WRN]` label.
    app.input_indicator = app.input_validity_verdict();
    let text = rendered_text(&mut app, &Theme::dark(), 80, 24);
    assert!(text.contains("[WRN]"), "the SQL DML warning surfaces as [WRN]:\n{text}");
}