WZhuo commented on issue #909: URL: https://github.com/apache/iceberg-cpp/issues/909#issuecomment-6971087881
I'd like to take this on. The repo now has enough fixtures to start: types, schema, partition transforms and reference tables. No other implementation has merged an integration yet. The existing PoCs for pyiceberg and iceberg-rust pin the fixtures as a git submodule, and I plan to follow that model: - Pin `apache/iceberg-verification` as a git submodule. A CMake variable can point to a different fixture directory. When the fixtures are absent, for example in a source tarball, the tests skip. - Add GTest parameterized suites, one test per `<subdir>/<case-id>`, that call the public APIs: `TypeFromJson`, `SchemaFromJson`, and `TransformFromString` → `Bind`. Only the subdirectories we opt into are run. - Keep a list of known failures in iceberg-cpp, with issue links. A listed case that starts passing fails the test, so the list stays current. I'd do this in phases: 1. Submodule, shared helpers, and the `types` and `schema` surfaces. 2. `partition-transforms`. 3. `reference-tables`: a full scan of every snapshot, including the v3 deletion vector and `initial-default` tables. This needs the `/tmp/iceberg-verification` symlink in CI, so Linux/macOS only for now. Any divergences we find will go back to iceberg-verification as feedback. Comments on the approach are welcome, especially submodule vs. `FetchContent`. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
