freakyzoidberg opened a new pull request, #10:
URL: https://github.com/apache/datasketches-tck/pull/10
## What
Move the Go snapshot pin from `ad31bdd` to `757a485` (the merge of
apache/datasketches-go#180) and regenerate `serialization/go/snapshots`.
## Why
Fixes the byte mismatch reported in apache/datasketches-go#185. At
`ad31bdd`, Go's `ToCompactSlice()` for HLL_6 and HLL_8 did not set the compact
flag, so preamble byte 5 was `0x00` where C++ and current Java write `0x08`.
apache/datasketches-go#180 fixed this, following
apache/datasketches-java@9a488f9.
`757a485` is used instead of Go `main` because the Go snapshot generator on
`main` currently fails in the theta and tuple packages: since
apache/datasketches-go#181, updates that theta legitimately screens out are
asserted as errors. With that fixed, `main` generates the same deterministic
snapshots as `757a485`.
## Snapshot changes
- **8 deterministic changes:** `hll6_n{1000,10000,100000,1000000}_go.sk` and
`hll8_n{1000,10000,100000,1000000}_go.sk` differ only at byte 5 (`0x00` →
`0x08`).
- **9 added:** `aos_multikey_n*_go.sk` and `aos_unicode_go.sk`. The Go
generator already produced these at the previous pin, but they were never added
when 9438efd removed the old `aos_multi_key_*` and `aos_unicode_strings` names.
The new names match the C++ and Java snapshots.
- **43 probabilistic snapshots regenerated:** bloom filter, REQ, reservoir
and VarOpt.
## Checks run locally
- `mise run tck -- snapshots check go`: the file set and all deterministic
snapshots reproduce; only probabilistic snapshots differ.
- `mise run check`: passes.
- datasketches-java `HllSketchCrossLanguageTest` (all test groups), using
the C++ and Go snapshots from this branch: 7/7 pass, so all 24 HLL snapshots
are byte-identical across C++, Go and Java. With the snapshots from
datasketches-java's currently pinned TCK revision, `checkBinaries2` and
`checkBinaries3` fail as reported in apache/datasketches-go#185.
## Follow-up
For datasketches-java's cross-language tests to pick this up, `TCK_REVISION`
in `tools/download_serialization_test_data.sh` (currently `d363b12`) needs to
move to this change once merged.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
--
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]
