test: consolidate write tests into single [[test]] binary (delta-io#2472)

## What changes are proposed in this pull request?

Collapses the 15 `kernel/tests/write_*.rs` files (split out in delta-io#2460)
into a single integration test binary declared via `[[test]]` in
`kernel/Cargo.toml`. Topic files move under `kernel/tests/write/` as
modules of `tests/write/main.rs`.


Test discovery and filtering are unchanged from a user POV: `cargo
nextest run --test write` runs the whole binary, `cargo nextest run
write::stats::` filters by topic.

`cargo test --test write_stats`-style compile pruning of one topic is
lost

## How was this change tested?
Measured locally, after `touch kernel/src/lib.rs`:

| | Binaries | `cargo build --tests -p delta_kernel --all-features` |
|---|---|---|
| before | 32 | 45.82s |
| after | 18 | 35.04s |

Author: DrakeLin

kernel/Cargo.toml

@@ -161,3 +161,12 @@ harness = false
 [[bench]]
 name = "expression_bench"
 harness = false
+
+# Consolidated write-path integration tests. Linking each integration test as a
+# separate binary is expensive (every binary re-links Arrow, Parquet, and the
+# default engine), so all write tests share a single binary whose entry point
+# is `tests/write/main.rs` and whose topics live as modules under that
+# directory.
+[[test]]
+name = "write"
+path = "tests/write/main.rs"

kernel/src/transaction/mod.rs

@@ -2121,7 +2121,7 @@ mod tests {
 
     // Note: Additional test coverage for partial file matching (where some files in a scan
     // have DV updates but others don't) is provided by the end-to-end integration test
-    // kernel/tests/dv.rs and kernel/tests/write_remove_dv.rs, which exercise
+    // kernel/tests/dv.rs and kernel/tests/write/remove_dv.rs, which exercise
     // the full deletion vector write workflow including the DvMatchVisitor logic.
 
     #[test]

kernel/tests/README.md

@@ -9,7 +9,7 @@ Test tables organized by feature area. Tables live in two locations:
 
 | Table | Location | Schema | Protocol (R/W) | Features | Description | Tests |
 |-------|----------|--------|----------|----------|-------------|-------|
-| `table-with-dv-small` | data/ | `value: int` | v3/v7 | r:`deletionVectors` w:`deletionVectors` | 10 rows, 2 soft-deleted by DV, 8 visible. Most heavily referenced test table. | `dv.rs::test_table_scan(with_dv)`, `write_remove_dv.rs::test_remove_files_adds_expected_entries`, `write_remove_dv.rs::test_update_deletion_vectors_adds_expected_entries`, `read.rs::with_predicate_and_removes`, `path.rs::test_to_uri/test_child/test_child_escapes`, `snapshot.rs::test_snapshot_read_metadata/test_new_snapshot/test_snapshot_new_from/test_read_table_with_missing_last_checkpoint/test_log_compaction_writer`, `deletion_vector.rs` tests, `transaction/mod.rs::setup_dv_enabled_table/test_add_files_schema/test_new_deletion_vector_path`, `default/parquet.rs` read test, `default/json.rs` read test, `log_compaction/tests.rs::create_mock_snapshot`, `resolve_dvs.rs` tests |
+| `table-with-dv-small` | data/ | `value: int` | v3/v7 | r:`deletionVectors` w:`deletionVectors` | 10 rows, 2 soft-deleted by DV, 8 visible. Most heavily referenced test table. | `dv.rs::test_table_scan(with_dv)`, `write::remove_dv::test_remove_files_adds_expected_entries`, `write::remove_dv::test_update_deletion_vectors_adds_expected_entries`, `read.rs::with_predicate_and_removes`, `path.rs::test_to_uri/test_child/test_child_escapes`, `snapshot.rs::test_snapshot_read_metadata/test_new_snapshot/test_snapshot_new_from/test_read_table_with_missing_last_checkpoint/test_log_compaction_writer`, `deletion_vector.rs` tests, `transaction/mod.rs::setup_dv_enabled_table/test_add_files_schema/test_new_deletion_vector_path`, `default/parquet.rs` read test, `default/json.rs` read test, `log_compaction/tests.rs::create_mock_snapshot`, `resolve_dvs.rs` tests |
 | `table-without-dv-small` | data/ | `value: long` | v1/v2 | | 10 rows, all visible. Companion to table-with-dv-small. | `dv.rs::test_table_scan(without_dv)`, `transaction/mod.rs::setup_non_dv_table/create_existing_table_txn/test_commit_io_error_returns_retryable_transaction`, `sequential_phase.rs::test_sequential_v2_with_commits_only/test_sequential_finish_before_exhaustion_error`, `parallel_phase.rs` tests, `scan/tests.rs::test_scan_metadata_paths/test_scan_metadata/test_scan_metadata_from_same_version` |
 | `with-short-dv` | data/ | `id: long, value: string, timestamp: timestamp, rand: double` | v3/v7 | r:`deletionVectors` w:`deletionVectors` | 2 files x 5 rows. First file has inline DV (`storageType="u"`) deleting 3 rows. | `read.rs::short_dv` |
 | `dv-partitioned-with-checkpoint` | golden_data/ | `value: int, part: int` partitioned by `part` | v3/v7 | r:`deletionVectors` w:`deletionVectors` | DVs on a partitioned table with a checkpoint | `golden_tables.rs::golden_test!` |
@@ -41,9 +41,9 @@ Test tables organized by feature area. Tables live in two locations:
 
 | Table | Location | Schema | Protocol (R/W) | Features | Description | Tests |
 |-------|----------|--------|----------|----------|-------------|-------|
-| `partition_cm/none` | data/ | `value: int, category: string` partitioned by `category` | v1/v1 | `columnMapping.mode=none` | Partitioned write with CM disabled | `write_column_mapping.rs::test_column_mapping_partitioned_write(cm_none)` |
-| `partition_cm/id` | data/ | `value: int, category: string` partitioned by `category` | v3/v7 | r:`columnMapping` w:`columnMapping`, `columnMapping.mode=id` | Partitioned write with CM id mode | `write_column_mapping.rs::test_column_mapping_partitioned_write(cm_id)` |
-| `partition_cm/name` | data/ | `value: int, category: string` partitioned by `category` | v3/v7 | r:`columnMapping` w:`columnMapping`, `columnMapping.mode=name` | Partitioned write with CM name mode | `write_column_mapping.rs::test_column_mapping_partitioned_write(cm_name)` |
+| `partition_cm/none` | data/ | `value: int, category: string` partitioned by `category` | v1/v1 | `columnMapping.mode=none` | Partitioned write with CM disabled | `write::column_mapping::test_column_mapping_partitioned_write(cm_none)` |
+| `partition_cm/id` | data/ | `value: int, category: string` partitioned by `category` | v3/v7 | r:`columnMapping` w:`columnMapping`, `columnMapping.mode=id` | Partitioned write with CM id mode | `write::column_mapping::test_column_mapping_partitioned_write(cm_id)` |
+| `partition_cm/name` | data/ | `value: int, category: string` partitioned by `category` | v3/v7 | r:`columnMapping` w:`columnMapping`, `columnMapping.mode=name` | Partitioned write with CM name mode | `write::column_mapping::test_column_mapping_partitioned_write(cm_name)` |
 | `table-with-columnmapping-mode-name` | golden_data/ | `ByteType: byte, ShortType: short, IntegerType: int, LongType: long, FloatType: float, DoubleType: double, decimal: decimal(10,2), BooleanType: boolean, StringType: string, BinaryType: binary, DateType: date, TimestampType: timestamp, nested_struct: struct{aa: string, ac: struct{aca: int}}, array_of_prims: array<int>, array_of_arrays: array<array<int>>, array_of_structs: array<struct{ab: long}>, map_of_prims: map<int,long>, map_of_rows: map<int,struct{ab: long}>, map_of_arrays: map<long,array<int>>` | v2/v5 | `columnMapping.mode=name` | Column mapping name mode | `golden_tables.rs::golden_test!` |
 | `table-with-columnmapping-mode-id` | golden_data/ | `ByteType: byte, ShortType: short, IntegerType: int, LongType: long, FloatType: float, DoubleType: double, decimal: decimal(10,2), BooleanType: boolean, StringType: string, BinaryType: binary, DateType: date, TimestampType: timestamp, nested_struct: struct{aa: string, ac: struct{aca: int}}, array_of_prims: array<int>, array_of_arrays: array<array<int>>, array_of_structs: array<struct{ab: long}>, map_of_prims: map<int,long>, map_of_rows: map<int,struct{ab: long}>, map_of_arrays: map<long,array<int>>` | v2/v5 | `columnMapping.mode=id` | Column mapping id mode | `golden_tables.rs::golden_test!` |
 
@@ -52,7 +52,7 @@ Test tables organized by feature area. Tables live in two locations:
 | Table | Location | Schema | Protocol (R/W) | Features | Description | Tests |
 |-------|----------|--------|----------|----------|-------------|-------|
 | `with_checkpoint_no_last_checkpoint` | data/ | `letter: string, int: long, date: date` | v1/v2 | `checkpointInterval=2` | Checkpoint at v2 but missing `_last_checkpoint` hint file | `snapshot.rs::test_read_table_with_checkpoint`, `scan/tests.rs::test_scan_with_checkpoint`, `sequential_phase.rs::test_sequential_checkpoint_no_commits`, `checkpoint_manifest.rs` tests, `sync/parquet.rs` test, `default/parquet.rs` test |
-| `external-table-different-nullability` | data/ | `i: int` | v1/v2 | `checkpointInterval=2` | Parquet files have different nullability than Delta schema; includes checkpoint | `write_clustered.rs::test_checkpoint_non_kernel_written_table` |
+| `external-table-different-nullability` | data/ | `i: int` | v1/v2 | `checkpointInterval=2` | Parquet files have different nullability than Delta schema; includes checkpoint | `write::clustered::test_checkpoint_non_kernel_written_table` |
 | `checkpoint` | golden_data/ | `intCol: int` | v1/v2 | | Basic checkpoint read | `golden_tables.rs::golden_test!(checkpoint_test)` |
 | `corrupted-last-checkpoint-kernel` | golden_data/ | `id: long` | v1/v2 | | Corrupted `_last_checkpoint` file | `golden_tables.rs::golden_test!` |
 | `multi-part-checkpoint` | golden_data/ | `id: long` | v1/v2 | `checkpointInterval=1` | Multi-part checkpoint files | `golden_tables.rs::golden_test!` |

kernel/tests/write_append.rs kernel/tests/write/append.rskernel/tests/write_append.rs renamed to kernel/tests/write/append.rs

@@ -19,9 +19,7 @@ use itertools::Itertools;
 use serde_json::{json, Deserializer};
 use test_utils::{create_table, engine_store_setup, set_json_value, setup_test_tables, test_read};
 
-mod common;
-
-use common::write_utils::{
+use crate::common::write_utils::{
     check_action_timestamps, get_and_check_all_parquet_sizes, get_simple_int_schema,
     validate_txn_id, write_data_and_check_result_and_stats, ZERO_UUID,
 };

kernel/tests/write_cdf.rs kernel/tests/write/cdf.rskernel/tests/write_cdf.rs renamed to kernel/tests/write/cdf.rs

@@ -17,9 +17,7 @@ use tempfile::{tempdir, TempDir};
 use test_utils::{assert_result_error_with_message, create_table, engine_store_setup};
 use url::Url;
 
-mod common;
-
-use common::write_utils::get_simple_int_schema;
+use crate::common::write_utils::get_simple_int_schema;
 
 // Helper function to create a table with CDF enabled
 async fn create_cdf_table(

kernel/tests/write_clustered.rs kernel/tests/write/clustered.rskernel/tests/write_clustered.rs renamed to kernel/tests/write/clustered.rs

@@ -22,9 +22,7 @@ use test_utils::{
 };
 use url::Url;
 
-mod common;
-
-use common::write_utils::{
+use crate::common::write_utils::{
     assert_column_mapping_mode, assert_min_max_stats, resolve_json_path, resolve_struct_field,
     set_table_properties,
 };

kernel/tests/write_column_mapping.rs kernel/tests/write/column_mapping.rskernel/tests/write_column_mapping.rs renamed to kernel/tests/write/column_mapping.rs

@@ -25,9 +25,7 @@ use test_utils::{
 };
 use url::Url;
 
-mod common;
-
-use common::write_utils::{
+use crate::common::write_utils::{
     assert_min_max_stats, get_parquet_field_id, resolve_struct_field, set_table_properties,
 };
 

kernel/tests/write_commit_info.rs kernel/tests/write/commit_info.rskernel/tests/write_commit_info.rs renamed to kernel/tests/write/commit_info.rs

@@ -14,9 +14,9 @@ use itertools::Itertools;
 use serde_json::{json, Deserializer};
 use test_utils::{set_json_value, setup_test_tables};
 
-mod common;
-
-use common::write_utils::{get_simple_int_schema, validate_timestamp, validate_txn_id, ZERO_UUID};
+use crate::common::write_utils::{
+    get_simple_int_schema, validate_timestamp, validate_txn_id, ZERO_UUID,
+};
 
 #[tokio::test]
 async fn test_commit_info() -> Result<(), Box<dyn std::error::Error>> {

kernel/tests/write_domain_metadata.rs kernel/tests/write/domain_metadata.rskernel/tests/write_domain_metadata.rs renamed to kernel/tests/write/domain_metadata.rs

@@ -8,9 +8,7 @@ use itertools::Itertools;
 use serde_json::Deserializer;
 use test_utils::{assert_result_error_with_message, create_table, engine_store_setup};
 
-mod common;
-
-use common::write_utils::get_simple_int_schema;
+use crate::common::write_utils::get_simple_int_schema;
 
 #[tokio::test]
 async fn test_set_domain_metadata_basic() -> Result<(), Box<dyn std::error::Error>> {

kernel/tests/write_ict.rs kernel/tests/write/ict.rskernel/tests/write_ict.rs renamed to kernel/tests/write/ict.rs

@@ -18,9 +18,7 @@ use tempfile::TempDir;
 use test_utils::engine_store_setup;
 use url::Url;
 
-mod common;
-
-use common::write_utils::get_simple_int_schema;
+use crate::common::write_utils::get_simple_int_schema;
 
 async fn get_ict_at_version(
     store: Arc<DynObjectStore>,