dpol1 commented on issue #3145: URL: https://github.com/apache/hugegraph/issues/3145#issuecomment-5779637268
@imbajin thanks for the references and the coverage correction. Confirmed: the corpus covers `encodeSortable`/`decodeSortable`, `encodeB64`/`decodeB64`, and `encodeSignedB64`/`decodeSignedB64`, never calls `encodeNumber`, and therefore pins nothing in `NumericUtil`. I corrected the matrix rows `LongEncoding`, `NumericUtil`, and `BytesBuffer` and the inventory in place, with a dated update block in the inventory, instead of adding another table. I ticked checkbox 2: #3146 is merged, and both tests run in the commons `UnitTestSuite`. @SebastianGruza thanks for the #3090 data point. I checked the store side: the `SelectIterator`/`FilterIterator` parse path exists, but no current traffic reaches it (no production caller selects properties; details in the inventory update). The `BytesBuffer` row now states that the byte-identical constraint binds today on the server encoder and its own readers and becomes a cross-process contract only if #3090 re-enables store-side filtering. For the storage rows of the matrix, I will cite the #3223 baselines rather than produce my own. Next step, towards the remaining checkbox (matrix reviewed): an RFC, filed as a child issue of #3110, that lets the maintainers decide whether and where a first Rust proof of concept in the Server is justified. It will cover: - the inventory candidates, plus at most one non-codec candidate that meets the same criteria; - three options side by side: an improved Java implementation, an in-process boundary, and an out-of-process boundary; - crossing costs measured on our own code: single and batched calls, copy versus read semantics, varying sizes; - the validation, packaging, and rollback sections the #3110 guardrails require. It presupposes neither a component nor an advantage; "not justified yet" is an acceptable conclusion, and the document will say whether the evidence is missing or contrary. Target: ready for review in about a month, before the discussion after 1.8.0. No Rust code enters this repository before that review. -- 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]
