From b8564b28fb7ba406c550a32ceff8e6bfb8301864 Mon Sep 17 00:00:00 2001 From: Andy Grove Date: Tue, 30 Jun 2026 13:02:37 -0600 Subject: [PATCH] ci: republish docs site when generated-doc sources change The docs deploy job only runs when compute-changes.py reports docs=true, which previously matched only .asf.yaml, the docs workflow, and docs/**. The published site also contains content generated by GenerateDocs (configs.md and the per-version expression compatibility pages), built from Scala sources. Changes to those sources did not republish the site unless a docs/ file happened to change in the same push, so generated pages could drift from main. Add the GenerateDocs input paths (CometConf, GenerateDocs, the serde and expressions packages, and the per-version shim directories) to the docs filter so the site republishes when generated content changes. --- dev/ci/compute-changes.py | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/dev/ci/compute-changes.py b/dev/ci/compute-changes.py index a708fde003a..01047e7dea7 100644 --- a/dev/ci/compute-changes.py +++ b/dev/ci/compute-changes.py @@ -80,6 +80,14 @@ ".asf.yaml", ".github/workflows/docs.yaml", "docs/**", + # Generated docs (configs.md, per-version expression compatibility pages) are + # built from these Scala sources by GenerateDocs, so changes to them must + # republish the site even when no docs/ file is touched. + "spark/src/main/scala/org/apache/comet/CometConf.scala", + "spark/src/main/scala/org/apache/comet/GenerateDocs.scala", + "spark/src/main/scala/org/apache/comet/serde/**", + "spark/src/main/scala/org/apache/comet/expressions/**", + "spark/src/main/spark-*/**", ], "spark_3_4": [ "native/**/src/**",