Repository navigation
refactor(sort): rename SpillCompress → SpillBlockCompress to disambiguate from CompressSpill - #557
Conversation
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: ASSERTIVE Plan: Pro Run ID: 📒 Files selected for processing (14)
WalkthroughThe spill compression stage is renamed from ChangesSpillBlockCompress stage
Estimated code review effort: 2 (Simple) | ~10 minutes Possibly related PRs
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Comment |
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## feat-runall #557 +/- ##
==============================================
Coverage ? 94.22%
==============================================
Files ? 111
Lines ? 50988
Branches ? 0
==============================================
Hits ? 48045
Misses ? 2943
Partials ? 0 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
897a269 to
8b5f1e7
Compare
8b5f1e7 to
d5e811b
Compare
d5e811b to
bdaaffe
Compare
bdaaffe to
161f50c
Compare
161f50c to
fbec2ab
Compare
fbec2ab to
51484c5
Compare
51484c5 to
b3d2de3
Compare
…guate from CompressSpill
Two distinct spill-compression steps had confusable word-swapped names:
- CompressSpill (chunk-granularity): compresses a whole sorted chunk AND
writes the spill file, in the P6 Phase-1 split
(SortBuffer -> CompressSpill -> SortSpillDecompress -> SortMerge).
- SpillCompress (block-granularity): compresses one raw SpillBlockEvent::Block
in place (no file write; SpillWrite writes), in the block-parallel split
(SpillGather -> SpillCompress -> SpillWrite).
They do different things, but the near-mirrored names invited confusion. Rename
the block-parallel step to SpillBlockCompress (and its module/file to
spill_block_compress), which also reads naturally against the SpillBlockCompressor
kernel it drives and the SpillBlockEvent::Block it consumes. Pure rename; no
behavior change.
b3d2de3 to
37b6451
Compare
|
@coderabbitai review |
✅ Action performedReview finished.
|
What
Two distinct spill-compression steps had confusable, word-swapped names:
CompressSpill(chunk-granularity) — compresses a whole sorted chunk and writes the spill file, in the P6 Phase-1 split (SortBuffer → CompressSpill → SortSpillDecompress → SortMerge).SpillCompress(block-granularity) — compresses one rawSpillBlockEvent::Blockin place (no file write;SpillWritewrites), in the block-parallel split (SpillGather → SpillCompress → SpillWrite).They do different things, but the near-mirrored names invited mix-ups. Rename the block-parallel step to
SpillBlockCompress(and its module/file tospill_block_compress), which reads naturally against theSpillBlockCompressorkernel it drives and theSpillBlockEvent::Blockit consumes.Notes
fgumi-pipeline-io/fgumi-sortare feat-runall-only, so nomain-reconcile conflict. Targetsnh/audit-3-parity.Verification
cargo ci-test2332 passed;cargo ci-lintclean.Summary by CodeRabbit
New Features
Documentation
Tests