mirror of https://github.com/hasura/graphql-engine.git synced 2024-12-15 17:31:56 +03:00

History

Samir Talwar 2c2ffa05bf All benchmarks must be Criterion benchmarks. (#498 ) This ensures only Criterion benchmarks get called by `cargo bench`. Previously, some tests were being loaded too. This is an issue because if we want to provide command-line arguments to Criterion, it will fail if we run a non-Criterion benchmark. We need to specify `harness = false` for each benchmark so that they don't get wrapped by Cargo, and `bench = false` for everything that doesn't contain benchmarks but _could_ (i.e. unit and integration tests). I find Cargo very strange in this regard. In addition, I fixed the validation benchmark by providing a valid SDL schema. V3_GIT_ORIGIN_REV_ID: 338ac5b1411eec7af32923863c5b6f3933c0454b		2024-04-23 13:00:24 +00:00
..
src	Add a binary for generating the OpenAPI schema for the metadata type. (#453 )	2024-04-10 07:36:55 +00:00
Cargo.toml	All benchmarks must be Criterion benchmarks. (#498 )	2024-04-23 13:00:24 +00:00
README.md	Add a binary for generating the OpenAPI schema for the metadata type. (#453 )	2024-04-10 07:36:55 +00:00

README.md

Metadata Schema Generator

When we make changes to the metadata schema, we need to regenerate the goldenfile to make the tests happy. We can use the following command to get the schema immediately for whatever reason:

$ cargo run --bin metadata-schema-generator