Ben Mehlow

This commit performs **maintenance** by **suppressing pre-existing `missing_brr` lint errors** across the codebase. It adds `brr_disable_reason: "NOT_YET_IMPLEMENTED"` to 96 `ci_sandcastle` targets within 69 BUCK files, including those in `fboss/agent` and `fboss/lib/platform_mapping_v2`. This **lint suppression** prevents new errors from blocking builds, with each entry annotated as `auto-derived` for future verification. The change **does not alter runtime behavior** but rather mitigates build-time linting issues.

This commit provides a **bug fix** for the **Base Revision Run (BRR) retry mechanism** that previously resulted in zero tests being run when the initial failure was recorded at a suite level (e.g., `"suite - main"`). The `get_filter()` logic is now updated to convert these specific suite-level names into prefix-based regular expressions (`^suite - `), ensuring that all individual tests within the affected suite are retried. This change prevents silent failures and improves the reliability of **test execution and reporting** for `run-as-bundle` and listing failures. New end-to-end tests have been added to `tests/e2e/test/test_brr_suite_level_retry.py` to verify this corrected behavior.

This commit introduces a **new capability** to **improve oncall propagation** within the **Buck2 test execution pipeline**, specifically addressing issues where Starlark macros fail to correctly pass oncall information. It ensures that the `buck.oncall` attribute defined in a BUCK file is used as a **default oncall for tests** if no explicit oncall is provided at the target level. This involves extending the **`ConfiguredTarget` data model** with a `package_oncall` field, updating **protobuf definitions**, and modifying the **test driver's logic** in `buck2_test` and `buck2_build_api` to propagate and apply this fallback. This **bug fix** for systemic oncall assignment problems significantly **reduces unowned builds and tests**, simplifying oncall management and improving ownership for internal projects.

This commit delivers a **bug fix** to enhance the **deduplication logic** for configurations and modes, specifically targeting **local `tpx` runs**. It addresses shortcomings in Buck's default deduplication by introducing explicit code to correctly handle multi-mode files, configs, and modifiers. This **maintenance** effort ensures accurate and efficient processing of build configurations, preventing potential issues from redundant or improperly managed settings within the `tpx` environment. New test cases in `tests/e2e/test/test_configs.py` were added to thoroughly verify the improved deduplication behavior and prevent regressions.

This commit **enhances the end-to-end test coverage** for **external configurations** by introducing new test cases and expanding existing ones in `tests/e2e/test/test_configs.py`. Specifically, it adds a test for multi-mode configuration files and extends validation to cover different targets. This **testing improvement** also includes **refactoring** a test helper function to accept a `target` parameter, thereby increasing test flexibility and maintainability. This work **improves the robustness** of the configuration system by ensuring broader scenario coverage and reducing potential regressions.

This commit introduces **logging for per-test external configurations** within the build system, specifically capturing the configuration after it has been enriched with target-specific modifiers and default modes. This **new capability** addresses the challenge of late availability of fully resolved configurations, providing an early point to inspect the final configuration state. The change primarily impacts the **configuration management** and **e2e testing subsystems**, as it directly **unblocks end-to-end tests** for features that rely on per-target configurations. This is validated by new test cases in `tests/e2e/test/test_configs.py` that assert correct default mode behavior and handle non-default mode files.

This commit **updates the warning message** within various `BUCK` files that are marked as unowned and suppressed with `# lint-ignore BUCKLINT missing-oncall-call-severe`. This **documentation update** significantly clarifies the severe consequences, explicitly stating that such files **will be deleted without notice** if an oncall is not added. The **maintenance change** aims to prevent developers from inadvertently losing their work by making the warning's severity undeniable. This impacts anyone interacting with these specific **unowned build files**, ensuring they understand the critical requirement to add an oncall before making changes.

This commit **enhances the parsing of command-line arguments** within the **`buck2_client_ctx` module** to correctly interpret short flags where the value is directly appended, such as `-cVALUE`. Specifically, the `get_representative_config_flags` function now properly handles single-string arguments like `'-c foo.bar=baz'`. This **new capability** ensures that generated **repro commands** accurately reflect and parse configurations, addressing a compatibility issue with existing usage patterns, including those in `fbcode//mode/libcxx`. The change improves the robustness of **CLI argument handling** for configuration flags within Buck2.

This commit **fixes a bug** introduced during a recent refactor, specifically within the **`buck2_test` module**. It **enables the `test` command** to correctly process multiple values for various configuration entries. Users can now specify multiple **config flags, files, modifiers, and target platforms** by appending them with a semicolon, restoring the intended flexibility and functionality for test configurations.

This commit **refactors the generation of config entry arguments** within the **`buck2_test` application**. It extracts the logic for creating these arguments into a new helper function, `generate_config_entry_args`, located in `app/buck2_test/src/command.rs`. This **refactoring** improves code organization and maintainability, and is accompanied by **new unit tests** to ensure the correctness of the argument generation. The added tests enhance the robustness and reliability of the `buck2_test` command's configuration handling.

This commit performs a **maintenance** task by adding lint-ignore comments to numerous `BUCK` files across various **build configurations** and **test suites**. Specifically, it **suppresses "missing oncall" warnings** in modules such as `docs/buck2_lab`, `integrations/rust-project`, and `prelude/apple`. This change is crucial to enable a related monitoring signal to be designated as **critical** without generating excessive noise from these non-actionable warnings, thereby improving the signal's effectiveness. The scope of this change is broad, affecting multiple `BUCK` files that previously triggered these specific lint warnings.

This commit **automates the assignment of oncall ownership** to `Buck TARGETS` files within the `fbcode` repository. It **adds `oncall` annotations** to these build configuration files, specifically assigning `rust_devx` to `integrations/rust-project/tests/BUCK` based on existing `contbuild` information. This is a **maintenance** task that enhances **ownership tracking** and improves **oncall routing** for various `fbcode` components. The change primarily impacts **Rust-related projects** by ensuring their build targets are correctly associated with the `rust_devx` oncall team.

add oncall annotation for TARGETS files in fbcode based on contbuild information - /tmp/oncalls_from_latest_version_of_script.txt

This commit **adds an `oncall` annotation** for `fboss_provisioning` to the `fboss/py/fboss/BUCK` file. This is a **chore** aimed at improving **ownership tracking** within the `fbcode` repository by leveraging `contbuild` information. The update ensures that the **Buck build system** has accurate metadata for oncall routing, specifically for the `fboss` module's provisioning targets. This enhances the maintainability and operational support for affected build configurations.

This commit **enhances the `buck2_test_runner` subsystem** by **forwarding maximum memory usage information** from the Buck build system to the test results. Specifically, the `runner.rs` module's `get_test_result` function is updated to include this new metric, which is then passed to downstream systems like TPX. This **new capability** provides more comprehensive data for test analysis, allowing for better insights into the resource consumption of executed tests. The change enriches the **test reporting subsystem** with valuable performance metrics.

Remove empty TARGETS and BUCK files

This commit introduces a **new capability** to track and report the **maximum memory used** during test execution. It **propagates this metric** from Buck's internal execution information by adding a `max_memory_used_bytes` field to key data structures like `TestResult` and `ExecutionResult2` across `buck2_test_api` and `buck2_test_proto`. This **data propagation** enables downstream systems, such as TPX, to receive and analyze memory consumption for individual tests, providing valuable performance insights.

This commit, the first in a three-part series, introduces the capability to pass the **interpreter host platform** information from Buck to **TPX** (TestX). It **refactors** the `buck2_interpreter_for_build` module by making the `host_info` method and `HostInfo` struct public, allowing the `buck2_test` command to retrieve this data. This **new capability** ensures that the host platform is consistently logged to Scuba for local test runs, laying the groundwork for improved test reproducibility. Ultimately, this information will be used by TestX to generate more accurate reproduction commands and instructions.

This commit introduces a **new capability** within the **`buck2_test` command** to process test configurations more effectively. It specifically extracts "mode files" from a set of representative configuration flags and passes them as a dedicated `'mode'` config entry to the test executor. This **feature enhancement** ensures that the test executor (`tpx`) receives precise, mode-specific configuration details, improving the accuracy and flexibility of test execution. This is the final part of a series aimed at providing comprehensive test configurations to `tpx`.

This commit introduces a **new capability** to forward representative test configuration flags from the Buck client to the server, enabling `tpx` to use them for generating accurate reproduction commands during test runs. It modifies the **`buck2_cli_proto`** to define a new `RepresentativeConfigFlag` message and updates **`buck2_client_ctx`** to extract and populate these flags from command-line arguments. The **`buck2_test`** module is then updated to extract and pass these flags from the client context to the test executor, ultimately improving the **reproducibility of test environments** by providing necessary configuration to `tpx`.