From a22454db06866ac0342783b76416c2ff9b086b51 Mon Sep 17 00:00:00 2001 From: Davide Angelocola Date: Sun, 26 Jul 2026 14:21:35 +0200 Subject: [PATCH] docs(changelog): record the ALP zero-length array fix Co-Authored-By: Claude Sonnet 5 --- CHANGELOG.md | 1 + 1 file changed, 1 insertion(+) diff --git a/CHANGELOG.md b/CHANGELOG.md index 6084fac1..5af82994 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -21,6 +21,7 @@ Real-world file-size work surfaced by the [Raincloud conformance corpus](https:/ ### Fixed - `InspectorTree`'s `usedEncodings` (surfaced by the CLI `inspect` command) now walks a Flat segment's full `ArrayNode` tree instead of only its root, so encodings nested inside another (e.g. `vortex.fsst` under a `vortex.masked` root) are no longer invisible. ([#298](https://github.com/dfa1/vortex-java/issues/298)) +- `AlpEncodingEncoder` no longer divides by zero on a zero-length `F64`/`F32` array — a real case on the Raincloud corpus, where the cascade's `Sparse` step can strip every value from an all-zero-bit chunk (e.g. a `congestion_surcharge` column that's `0.0` for a whole chunk). ([872b0554](https://github.com/dfa1/vortex-java/commit/872b0554)) ### Performance