diff --git a/.github/workflows/build_and_test.yml b/.github/workflows/build_and_test.yml
new file mode 100644
index 0000000000..f58d87d112
--- /dev/null
+++ b/.github/workflows/build_and_test.yml
@@ -0,0 +1,38 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements.  See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership.  The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License.  You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied.  See the License for the
+# specific language governing permissions and limitations
+# under the License.
+
+# Reusable workflow invoked by build_main.yml. Modelled on Apache Spark's
+# build_and_test.yml: the heavy work runs wherever the calling push lands,
+# i.e. on the contributor's fork. Delegates the test matrix to the umbrella
+# ci.yml workflow, which fans out to the per-Spark-version reusables.
+
+name: Build and test
+
+on:
+  workflow_call:
+
+jobs:
+  ci:
+    name: CI
+    uses: ./.github/workflows/ci.yml
+    secrets: inherit
+    with:
+      # Force every gated job to run on fork-CI: contributor pushes don't necessarily
+      # touch paths that match ci.yml's filters (e.g. when iterating on the fork-CI
+      # workflow files themselves), so honour the Spark-style "run the full matrix"
+      # contract instead of skipping silently.
+      force_all: true
diff --git a/.github/workflows/build_main.yml b/.github/workflows/build_main.yml
new file mode 100644
index 0000000000..dcc0f41ef6
--- /dev/null
+++ b/.github/workflows/build_main.yml
@@ -0,0 +1,35 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements.  See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership.  The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License.  You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied.  See the License for the
+# specific language governing permissions and limitations
+# under the License.
+
+# Modelled on Apache Spark's build_main.yml. Triggers on push to any branch,
+# which means contributor pushes to a fork run CI on the fork's Actions
+# minutes/runners rather than on apache/datafusion-comet. The bridge to the
+# upstream PR check is notify_test_workflow.yml.
+
+name: "Build"
+
+on:
+  push:
+    branches:
+    - '**'
+
+jobs:
+  call-build-and-test:
+    permissions:
+      packages: write
+    name: Run
+    uses: ./.github/workflows/build_and_test.yml
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index b5c13261c2..e50cb6b6f6 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -22,16 +22,30 @@
 name: CI
 
 concurrency:
-  group: ${{ github.repository }}-${{ github.head_ref || github.sha }}-${{ github.workflow }}
+  # Use github.ref (branch ref) instead of github.sha for the push-event fallback
+  # so consecutive pushes to the same branch cancel each other; sha would make
+  # every push a new group and stack runs up.
+  group: ${{ github.repository }}-${{ github.head_ref || github.ref }}-${{ github.workflow }}
   cancel-in-progress: true
 
+# CI DISABLED 2026-05-31 — auto-triggers commented out. Active triggers:
+#   * workflow_dispatch — manual runs
+#   * workflow_call     — invoked by build_and_test.yml on contributor forks
+# To rollback: restore the commented `pull_request` and `push` blocks below.
 on:
-  pull_request:
-    types: [opened, synchronize, reopened, labeled]
-  push:
-    branches:
-      - main
   workflow_dispatch:
+  workflow_call:
+    inputs:
+      force_all:
+        description: 'Force-enable every gated job (used by fork-CI so the full matrix runs regardless of which paths changed).'
+        required: false
+        type: boolean
+        default: false
+  # pull_request:
+  #   types: [opened, synchronize, reopened, labeled]
+  # push:
+  #   branches:
+  #     - main
 
 jobs:
   # ---------------------------------------------------------------------------
@@ -43,7 +57,7 @@ jobs:
   # ---------------------------------------------------------------------------
   preflight:
     name: Preflight
-    runs-on: ubuntu-slim
+    runs-on: ${{ github.repository == 'apache/datafusion-comet' && 'ubuntu-slim' || 'ubuntu-24.04' }}
     steps:
       - uses: actions/checkout@v6
 
@@ -89,7 +103,7 @@ jobs:
   changes:
     name: Detect changes
     needs: preflight
-    runs-on: ubuntu-slim
+    runs-on: ${{ github.repository == 'apache/datafusion-comet' && 'ubuntu-slim' || 'ubuntu-24.04' }}
     outputs:
       build_linux:  ${{ steps.compute.outputs.build_linux }}
       build_macos:  ${{ steps.compute.outputs.build_macos }}
@@ -118,9 +132,10 @@ jobs:
           PR_HEAD_SHA: ${{ github.event.pull_request.head.sha }}
           PUSH_BEFORE: ${{ github.event.before }}
           PUSH_AFTER:  ${{ github.sha }}
+          FORCE_ALL:   ${{ inputs.force_all }}
         run: |
           set -euo pipefail
-          if [[ "$EVENT_NAME" == "workflow_dispatch" ]]; then
+          if [[ "$EVENT_NAME" == "workflow_dispatch" || "$FORCE_ALL" == "true" ]]; then
             for key in build_linux build_macos benchmark docs spark_3_4 spark_3_5 spark_4_0 spark_4_1 iceberg_1_8 iceberg_1_9 iceberg_1_10; do
               echo "${key}=true" >> "$GITHUB_OUTPUT"
             done
diff --git a/.github/workflows/codeql.yml b/.github/workflows/codeql.yml
index 76c166bc3b..0cc8f376c2 100644
--- a/.github/workflows/codeql.yml
+++ b/.github/workflows/codeql.yml
@@ -22,13 +22,16 @@ concurrency:
   group: ${{ github.repository }}-${{ github.head_ref || github.sha }}-${{ github.workflow }}
   cancel-in-progress: true
 
+# CI DISABLED 2026-05-31 — auto-triggers commented out, only manual workflow_dispatch remains.
+# To rollback: remove `workflow_dispatch:` and restore the commented `push`/`pull_request`/`schedule` blocks below.
 on:
-  push:
-    branches: [ "main" ]
-  pull_request:
-    branches: [ "main" ]
-  schedule:
-    - cron: '16 4 * * 1'
+  workflow_dispatch:
+  # push:
+  #   branches: [ "main" ]
+  # pull_request:
+  #   branches: [ "main" ]
+  # schedule:
+  #   - cron: '16 4 * * 1'
 
 permissions:
   contents: read
diff --git a/.github/workflows/images/workflow-enable-button.png b/.github/workflows/images/workflow-enable-button.png
new file mode 100644
index 0000000000..f7299f233a
Binary files /dev/null and b/.github/workflows/images/workflow-enable-button.png differ
diff --git a/.github/workflows/label_new_issues.yml b/.github/workflows/label_new_issues.yml
index bbb4fb150d..01e263d0e2 100644
--- a/.github/workflows/label_new_issues.yml
+++ b/.github/workflows/label_new_issues.yml
@@ -17,9 +17,12 @@
 
 name: Label new issues with requires-triage
 
+# CI DISABLED 2026-05-31 — auto-trigger commented out, only manual workflow_dispatch remains.
+# To rollback: remove `workflow_dispatch:` and restore the commented `issues` block below.
 on:
-  issues:
-    types: [opened]
+  workflow_dispatch:
+  # issues:
+  #   types: [opened]
 
 permissions:
   issues: write
diff --git a/.github/workflows/miri.yml b/.github/workflows/miri.yml
index ad318888c3..7d711225e3 100644
--- a/.github/workflows/miri.yml
+++ b/.github/workflows/miri.yml
@@ -21,10 +21,12 @@ concurrency:
   group: ${{ github.repository }}-${{ github.head_ref || github.sha }}-${{ github.workflow }}
   cancel-in-progress: true
 
+# CI DISABLED 2026-05-31 — nightly schedule commented out, only manual workflow_dispatch remains.
+# To rollback: restore the commented `schedule` block below.
 on:
   # nightly safety check
-  schedule:
-    - cron: '0 4 * * *'
+  # schedule:
+  #   - cron: '0 4 * * *'
   # manual trigger
   # https://docs.github.com/en/actions/managing-workflow-runs/manually-running-a-workflow
   workflow_dispatch:
diff --git a/.github/workflows/notify_test_workflow.yml b/.github/workflows/notify_test_workflow.yml
new file mode 100644
index 0000000000..99be466956
--- /dev/null
+++ b/.github/workflows/notify_test_workflow.yml
@@ -0,0 +1,171 @@
+#
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements.  See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership.  The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License.  You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied.  See the License for the
+# specific language governing permissions and limitations
+# under the License.
+
+# Intentionally has a general name.
+# because the test status check created in GitHub Actions
+# currently randomly picks any associated workflow.
+# So, the name was changed to make sense in that context too.
+# See also https://github.community/t/specify-check-suite-when-creating-a-checkrun/118380/10
+name: On pull request update
+on:
+  pull_request_target:
+    types: [opened, reopened, synchronize]
+
+jobs:
+  notify:
+    name: Notify test workflow
+    runs-on: ubuntu-slim
+    permissions:
+      actions: read
+      checks: write
+    steps:
+      - name: "Notify test workflow"
+        uses: actions/github-script@v9
+        with:
+          github-token: ${{ secrets.GITHUB_TOKEN }}
+          script: |
+            const endpoint = 'GET /repos/:owner/:repo/actions/workflows/:id/runs?&branch=:branch'
+            const check_run_endpoint = 'GET /repos/:owner/:repo/commits/:ref/check-runs?per_page=100'
+
+            // TODO: Should use pull_request.user and pull_request.user.repos_url?
+            // If a different person creates a commit to another forked repo,
+            // it wouldn't be able to detect.
+            const params = {
+              owner: context.payload.pull_request.head.repo.owner.login,
+              repo: context.payload.pull_request.head.repo.name,
+              id: 'build_main.yml',
+              branch: context.payload.pull_request.head.ref,
+            }
+            const check_run_params = {
+              owner: context.payload.pull_request.head.repo.owner.login,
+              repo: context.payload.pull_request.head.repo.name,
+              ref: context.payload.pull_request.head.ref,
+            }
+
+            console.log('Ref: ' + context.payload.pull_request.head.ref)
+            console.log('SHA: ' + context.payload.pull_request.head.sha)
+
+            // Wait 3 seconds to make sure the fork repository triggered a workflow.
+            await new Promise(r => setTimeout(r, 3000))
+
+            let runs
+            try {
+              runs = await github.request(endpoint, params)
+            } catch (error) {
+              console.error(error)
+              // Assume that runs were not found.
+            }
+
+            const name = 'Build'
+            const head_sha = context.payload.pull_request.head.sha
+            let status = 'queued'
+
+            if (!runs || runs.data.workflow_runs.length === 0) {
+              status = 'completed'
+              const conclusion = 'action_required'
+
+              github.rest.checks.create({
+                owner: context.repo.owner,
+                repo: context.repo.repo,
+                name: name,
+                head_sha: head_sha,
+                status: status,
+                conclusion: conclusion,
+                output: {
+                  title: 'Workflow run detection failed',
+                  summary: `
+            Unable to detect the workflow run for testing the changes in your PR.
+
+            1. If you did not enable GitHub Actions in your forked repository, please enable it by clicking the button as shown in the image below. See also [Managing Github Actions Settings for a repository](https://docs.github.com/en/repositories/managing-your-repositorys-settings-and-features/enabling-features-for-your-repository/managing-github-actions-settings-for-a-repository) for more details.
+            2. It is possible your branch is based on the old \`main\` branch in Apache Datafusion Comet, please sync your branch to the latest main branch. For example as below:
+                \`\`\`bash
+                git fetch upstream
+                git rebase upstream/main
+                git push origin YOUR_BRANCH --force
+                \`\`\``,
+                  images: [
+                    {
+                      alt: 'enabling workflows button',
+                      image_url: 'https://raw.githubusercontent.com/apache/datafusion-comet/main/.github/workflows/images/workflow-enable-button.png'
+                    }
+                  ]
+                }
+              })
+            } else {
+              const run_id = runs.data.workflow_runs[0].id
+
+              if (runs.data.workflow_runs[0].head_sha != context.payload.pull_request.head.sha) {
+                throw new Error('There was a new unsynced commit pushed. Please retrigger the workflow.');
+              }
+
+              // Here we get check run ID to provide Check run view instead of Actions view, see also SPARK-37879.
+              let retryCount = 0;
+              let check_run_head;
+              while (retryCount < 3) {
+                const check_runs = await github.request(check_run_endpoint, check_run_params);
+                check_run_head = check_runs.data.check_runs.find(r => r.name === "Run / CI / Preflight");
+                if (check_run_head) {
+                  break;
+                }
+                retryCount++;
+                if (retryCount < 3) {
+                  await new Promise(resolve => setTimeout(resolve, 3000));
+                }
+              }
+              if (!check_run_head) {
+                throw new Error('Failed to retrieve check_run_head after 3 attempts');
+              }
+
+              if (check_run_head.head_sha != context.payload.pull_request.head.sha) {
+                throw new Error('There was a new unsynced commit pushed. Please retrigger the workflow.');
+              }
+
+              const check_run_url = 'https://github.com/'
+                + context.payload.pull_request.head.repo.full_name
+                + '/runs/'
+                + check_run_head.id
+              console.log('Check run URL: ' + check_run_url)
+
+              const actions_url = 'https://github.com/'
+                + context.payload.pull_request.head.repo.full_name
+                + '/actions/runs/'
+                + run_id
+              console.log('Actions URL: ' + actions_url)
+
+              github.rest.checks.create({
+                owner: context.repo.owner,
+                repo: context.repo.repo,
+                name: name,
+                head_sha: head_sha,
+                status: status,
+                output: {
+                  title: 'Test results',
+                  summary: '[See test results](' + check_run_url + ')\n\n'
+                    + 'If the tests fail for reasons unrelated to this pull request, '
+                    + 'please rerun the workflow in your forked repository.\n'
+                    + 'If the failures are related to this pull request, '
+                    + 'please investigate them and push follow-up changes.',
+                  text: JSON.stringify({
+                    owner: context.payload.pull_request.head.repo.owner.login,
+                    repo: context.payload.pull_request.head.repo.name,
+                    run_id: run_id
+                  })
+                },
+                details_url: actions_url,
+              })
+            }
\ No newline at end of file
diff --git a/.github/workflows/pr_title_check.yml b/.github/workflows/pr_title_check.yml
index fe3674acb4..46e97df5c5 100644
--- a/.github/workflows/pr_title_check.yml
+++ b/.github/workflows/pr_title_check.yml
@@ -21,9 +21,12 @@ concurrency:
   group: ${{ github.repository }}-${{ github.head_ref || github.sha }}-${{ github.workflow }}
   cancel-in-progress: true
 
+# CI DISABLED 2026-05-31 — auto-trigger commented out, only manual workflow_dispatch remains.
+# To rollback: remove `workflow_dispatch:` and restore the commented `pull_request` block below.
 on:
-  pull_request:
-    types: [opened, edited, reopened]
+  workflow_dispatch:
+  # pull_request:
+  #   types: [opened, edited, reopened]
 
 jobs:
   check-pr-title:
diff --git a/.github/workflows/spark_sql_test_reusable.yml b/.github/workflows/spark_sql_test_reusable.yml
index 4e89cda860..610806534d 100644
--- a/.github/workflows/spark_sql_test_reusable.yml
+++ b/.github/workflows/spark_sql_test_reusable.yml
@@ -121,6 +121,21 @@ jobs:
       image: amd64/rust
     steps:
       - uses: actions/checkout@v6
+      - name: Free up disk space (container)
+        # Mirror apache/spark's dev/free_disk_space_container. The amd64/rust
+        # container inherits GitHub's tool-cache via the /__t mount; strip the
+        # entries we never use (CodeQL, Go, Node) to free ~5-10 GB before the
+        # Spark clone, Maven dep tree, sbt build cache, and per-test-suite
+        # warehouse dirs all start fighting for the runner's disk.
+        shell: bash
+        run: |
+          echo "Disk usage before cleanup:"
+          df -h || true
+          rm -rf /__t/CodeQL || true
+          rm -rf /__t/go     || true
+          rm -rf /__t/node   || true
+          echo "Disk usage after cleanup:"
+          df -h || true
       - name: Setup Rust & Java toolchain
         uses: ./.github/actions/setup-builder
         with:
@@ -152,9 +167,10 @@ jobs:
           if [ "${{ inputs.spark-short }}" != "4.0" ] || [ "${{ inputs.java }}" != "21" ]; then
             export SERIAL_SBT_TESTS=1
           fi
-          # Cap parallel forked test JVMs at 1 so that even when
-          # SparkParallelTestGrouping is enabled we don't blow the
-          # 7 GB runner budget (each forked test JVM has -Xmx2g).
+          # Cap parallel forked test JVMs at 1: on a 7 GB runner the 3 GB SBT
+          # controller and a single 5 GB forked test JVM (HEAP_SIZE override
+          # below) already over-commit RAM; any concurrency on top would push
+          # us straight into OOMs.
           NOLINT_ON_COMPILE=true ENABLE_COMET=true ENABLE_COMET_ONHEAP=true ENABLE_COMET_LOG_FALLBACK_REASONS=${{ inputs.collect-fallback-logs }} \
             build/sbt -Dsbt.log.noformat=true -mem $SBT_MEM \
               'set Global / concurrentRestrictions := Seq(Tags.limit(Tags.ForkedTestGroup, 1))' \
@@ -164,17 +180,35 @@ jobs:
           fi
         env:
           LC_ALL: "C.UTF-8"
-          # Standard GitHub runners have 7 GB RAM; cap SBT heap so forked test
-          # JVMs fit alongside it.
           SBT_MEM: "3072"
-          # Mirror Spark's own JDK 21 / 25 CI workaround. apache/spark's
-          # build_java21.yml and build_java25.yml set this same env var to
-          # process-isolate the V1/V2 Parquet and Orc source suites because
-          # they exhibit cross-suite resource interactions (file-stream and
-          # thread leaks) under the newer JDKs. project/SparkBuild.scala
-          # reads DEDICATED_JVM_SBT_TESTS and forks a separate JVM per
-          # listed suite. Empty value is a safe no-op.
-          DEDICATED_JVM_SBT_TESTS: ${{ inputs.spark-short == '4.0' && 'org.apache.spark.sql.execution.datasources.parquet.ParquetFileFormatV1Suite,org.apache.spark.sql.execution.datasources.parquet.ParquetFileFormatV2Suite,org.apache.spark.sql.execution.datasources.orc.OrcSourceV1Suite,org.apache.spark.sql.execution.datasources.orc.OrcSourceV2Suite' || '' }}
+          HEAP_SIZE: "3g"
+          # Cap forked test JVM direct memory. Without this, -XX:MaxDirectMemorySize
+          # defaults to ≈ -Xmx (5 GB), letting Netty / DirectByteBuffer / shuffle
+          # buffers grow alongside the heap and push host RSS past the 7 GB
+          # runner ceiling. JAVA_TOOL_OPTIONS is honoured by every child JVM;
+          # SBT picks it up too but allocates negligible direct memory itself.
+          JAVA_TOOL_OPTIONS: "-XX:MaxDirectMemorySize=1g"
+          # Force these suites into their own forked JVMs so per-suite state
+          # (codegen classes, native pools, shuffle buffers) is reclaimed on
+          # JVM exit instead of accumulating across a single testOnly run:
+          #   * AdaptiveQueryExecSuite — long, codegen- and shuffle-heavy;
+          #     observed to SIGKILL the host on 7 GB runners after ~10 tests.
+          #   * Parquet/Orc V1/V2 source suites — Spark 4.0 only; mirrors
+          #     apache/spark's build_java21.yml / build_java25.yml workaround
+          #     for cross-suite file-stream / thread leaks under newer JDKs.
+          # SparkBuild.scala reads this env var and installs the testGrouping
+          # regardless of SERIAL_SBT_TESTS, so listed suites still get their
+          # own JVM in serial-test mode (only run policy changes).
+          DEDICATED_JVM_SBT_TESTS: ${{ inputs.spark-short == '4.0' && 'org.apache.spark.sql.execution.adaptive.AdaptiveQueryExecSuite,org.apache.spark.sql.execution.datasources.parquet.ParquetFileFormatV1Suite,org.apache.spark.sql.execution.datasources.parquet.ParquetFileFormatV2Suite,org.apache.spark.sql.execution.datasources.orc.OrcSourceV1Suite,org.apache.spark.sql.execution.datasources.orc.OrcSourceV2Suite' || 'org.apache.spark.sql.execution.adaptive.AdaptiveQueryExecSuite' }}
+      - name: Dump kernel + memory state on failure
+        if: failure()
+        shell: bash
+        run: |
+          echo '== /proc/meminfo =='; cat /proc/meminfo 2>/dev/null || true
+          echo '== /proc/swaps ==';   cat /proc/swaps   2>/dev/null || true
+          echo '== top RSS ==';       ps -eo pid,rss,cmd --sort=-rss 2>/dev/null | head -20 || true
+          echo '== dmesg (best effort) ==';
+          dmesg -T 2>/dev/null | tail -200 || cat /var/log/kern.log 2>/dev/null | tail -200 || echo 'kernel log unavailable inside container'
       - name: Upload fallback log
         if: ${{ inputs.collect-fallback-logs }}
         uses: actions/upload-artifact@v7
diff --git a/.github/workflows/stale.yml b/.github/workflows/stale.yml
index 7ad3a4c011..6010856602 100644
--- a/.github/workflows/stale.yml
+++ b/.github/workflows/stale.yml
@@ -16,9 +16,12 @@
 # under the License.
 
 name: "Close stale PRs"
+# CI DISABLED 2026-05-31 — schedule commented out, only manual workflow_dispatch remains.
+# To rollback: remove `workflow_dispatch:` and restore the commented `schedule` block below.
 on:
-  schedule:
-    - cron: "30 1 * * *"
+  workflow_dispatch:
+  # schedule:
+  #   - cron: "30 1 * * *"
 
 jobs:
   close-stale-prs:
diff --git a/.github/workflows/take.yml b/.github/workflows/take.yml
index 9e84da8da6..45ae416d64 100644
--- a/.github/workflows/take.yml
+++ b/.github/workflows/take.yml
@@ -16,9 +16,12 @@
 # under the License.
 
 name: Assign/unassign the issue via `take` or `untake` comment
+# CI DISABLED 2026-05-31 — auto-trigger commented out, only manual workflow_dispatch remains.
+# To rollback: remove `workflow_dispatch:` and restore the commented `issue_comment` block below.
 on:
-  issue_comment:
-    types: created
+  workflow_dispatch:
+  # issue_comment:
+  #   types: created
 
 permissions:
   issues: write
diff --git a/.github/workflows/update_build_status.yml b/.github/workflows/update_build_status.yml
new file mode 100644
index 0000000000..c7c66ccb08
--- /dev/null
+++ b/.github/workflows/update_build_status.yml
@@ -0,0 +1,111 @@
+#
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements.  See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership.  The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License.  You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied.  See the License for the
+# specific language governing permissions and limitations
+# under the License.
+#
+
+name: Update build status workflow
+
+on:
+  schedule:
+    - cron: "*/15 * * * *"
+
+jobs:
+  update:
+    name: Update build status
+    runs-on: ubuntu-22.04
+    permissions:
+      actions: read
+      checks: write
+    steps:
+      - name: "Update build status"
+        uses: actions/github-script@v9
+        with:
+          github-token: ${{ secrets.GITHUB_TOKEN }}
+          script: |
+            const endpoint = 'GET /repos/:owner/:repo/pulls?state=:state'
+            const params = {
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              state: 'open'
+            }
+
+            // See https://docs.github.com/en/graphql/reference/enums#mergestatestatus
+            const maybeReady = ['behind', 'clean', 'draft', 'has_hooks', 'unknown', 'unstable'];
+
+            // Iterate open PRs
+            for await (const prs of github.paginate.iterator(endpoint,params)) {
+              // Each page
+              for await (const pr of prs.data) {
+                console.log('SHA: ' + pr.head.sha)
+                console.log('  Mergeable status: ' + pr.mergeable_state)
+                if (pr.mergeable_state == null || maybeReady.includes(pr.mergeable_state)) {
+                  const checkRuns = await github.request('GET /repos/{owner}/{repo}/commits/{ref}/check-runs', {
+                    owner: context.repo.owner,
+                    repo: context.repo.repo,
+                    ref: pr.head.sha
+                  })
+
+                  // Iterate GitHub Checks in the PR
+                  for await (const cr of checkRuns.data.check_runs) {
+                    if (cr.name == 'Build' && cr.conclusion != "action_required") {
+                      // text contains parameters to make request in JSON.
+                      const params = JSON.parse(cr.output.text)
+
+                      // Get the workflow run in the forked repository
+                      let run
+                      try {
+                        run = await github.request('GET /repos/{owner}/{repo}/actions/runs/{run_id}', params)
+                      } catch (error) {
+                        console.error(error)
+                        // Run not found. This can happen when the PR author removes GitHub Actions runs or
+                        // disables GitHub Actions.
+                        continue
+                      }
+
+                      // Keep syncing the status of the checks
+                      if (run.data.status == 'completed') {
+                        console.log('    Run ' + cr.id + ': set status (' + run.data.status + ') and conclusion (' + run.data.conclusion + ')')
+                        const response = await github.request('PATCH /repos/{owner}/{repo}/check-runs/{check_run_id}', {
+                          owner: context.repo.owner,
+                          repo: context.repo.repo,
+                          check_run_id: cr.id,
+                          output: cr.output,
+                          status: run.data.status,
+                          conclusion: run.data.conclusion,
+                          details_url: run.data.details_url
+                        })
+                      } else {
+                        // PATCH /check-runs accepts only queued | in_progress | completed,
+                        // but workflow_run may also report requested / waiting / pending.
+                        const status = run.data.status == 'in_progress' ? 'in_progress' : 'queued'
+                        console.log('    Run ' + cr.id + ': set status (' + run.data.status + ' -> ' + status + ')')
+                        const response = await github.request('PATCH /repos/{owner}/{repo}/check-runs/{check_run_id}', {
+                          owner: context.repo.owner,
+                          repo: context.repo.repo,
+                          check_run_id: cr.id,
+                          output: cr.output,
+                          status: status,
+                          details_url: run.data.details_url
+                        })
+                      }
+
+                      break
+                    }
+                  }
+                }
+              }
+            }
\ No newline at end of file
diff --git a/dev/diffs/4.1.2.diff b/dev/diffs/4.1.2.diff
index f9911ce56d..580d0bcd8f 100644
--- a/dev/diffs/4.1.2.diff
+++ b/dev/diffs/4.1.2.diff
@@ -1662,6 +1662,28 @@ index ede5d285932..c9a8abb5a94 100644
    import testImplicits._
  
    before {
+diff --git a/sql/core/src/test/scala/org/apache/spark/sql/errors/QueryExecutionAnsiErrorsSuite.scala b/sql/core/src/test/scala/org/apache/spark/sql/errors/QueryExecutionAnsiErrorsSuite.scala
+index 64bb5a289b3..13fc3c6b0ef 100644
+--- a/sql/core/src/test/scala/org/apache/spark/sql/errors/QueryExecutionAnsiErrorsSuite.scala
++++ b/sql/core/src/test/scala/org/apache/spark/sql/errors/QueryExecutionAnsiErrorsSuite.scala
+@@ -20,6 +20,7 @@ import org.apache.spark._
+ import org.apache.spark.SparkBuildInfo
+ import org.apache.spark.scheduler.{SparkListener, SparkListenerTaskStart}
+ import org.apache.spark.sql.QueryTest
++import org.apache.spark.sql.IgnoreComet
+ import org.apache.spark.sql.catalyst.expressions.{CaseWhen, Cast, CheckOverflowInTableInsert, ExpressionProxy, Literal, SubExprEvaluationRuntime}
+ import org.apache.spark.sql.catalyst.plans.logical.OneRowRelation
+ import org.apache.spark.sql.classic.SparkSession
+@@ -286,7 +287,8 @@ class QueryExecutionAnsiErrorsSuite extends QueryTest
+     )
+   }
+ 
+-  test("INVALID_DATETIME_PATTERN with non-constant pattern") {
++  test("INVALID_DATETIME_PATTERN with non-constant pattern",
++    IgnoreComet("Comet exception type mismatch")) {
+     withTable("patterns") {
+       sql("create table patterns(pattern string) using parquet")
+       sql("insert into patterns values ('yyyyMMddHHMIss')")
 diff --git a/sql/core/src/test/scala/org/apache/spark/sql/errors/QueryExecutionErrorsSuite.scala b/sql/core/src/test/scala/org/apache/spark/sql/errors/QueryExecutionErrorsSuite.scala
 index fcecaf25d4c..e5a511022cc 100644
 --- a/sql/core/src/test/scala/org/apache/spark/sql/errors/QueryExecutionErrorsSuite.scala
diff --git a/docs/source/contributor-guide/development.md b/docs/source/contributor-guide/development.md
index 3785358cf3..31dff866b0 100644
--- a/docs/source/contributor-guide/development.md
+++ b/docs/source/contributor-guide/development.md
@@ -660,6 +660,52 @@ make          # Build everything and update generated docs
 make test     # Run tests (optional but recommended)
 ```
 
+## Forked-Branch CI
+
+Comet runs PR CI on the contributor's fork rather than on `apache/datafusion-comet`. The heavy build/test matrix consumes the fork owner's GitHub Actions minutes and runners; the upstream repo only hosts a thin `Build` check that links to the fork's run. The design mirrors Apache Spark's fork-CI bridge.
+
+### Flow
+
+```mermaid
+flowchart TD
+    subgraph fork["Forked repo (contributor)"]
+        P([push to any branch]) --> BM["build_main.yml — job: Run"]
+        BM -->|workflow_call| BAT["build_and_test.yml (reusable):<br/>ci.yml fan-out, force_all=true"]
+        BAT --> CR["check runs:<br/>Run / CI / Preflight, per-Spark-version matrix, ..."]
+    end
+
+    subgraph up["Upstream repo (apache/datafusion-comet)"]
+        PRT([pull_request_target]) --> N["notify_test_workflow.yml"]
+        SCH([schedule: every 15 min]) --> U["update_build_status.yml"]
+        N -->|create| B(["Build check on PR"])
+        U -->|"PATCH status / conclusion"| B
+    end
+
+    BM -.->|"① find run (id: build_main.yml)"| N
+    CR -.->|"② link check-run view"| N
+    BM -.->|"③ poll run status"| U
+```
+
+### Workflow responsibilities
+
+- **`build_main.yml`** (fork) — triggers on `push` to any branch and calls `build_and_test.yml`. Contributor pushes to their fork run here, on the fork's runners.
+- **`build_and_test.yml`** (fork, reusable) — invokes `ci.yml` with `force_all: true`, so the full per-Spark-version matrix runs even when changed paths don't match `ci.yml`'s filters (e.g. when iterating on the workflow files themselves).
+- **`notify_test_workflow.yml`** (upstream) — runs on `pull_request_target` for opened/reopened/synchronize. Looks up the matching `build_main.yml` run on the PR head's fork+branch, then creates a `Build` check on the PR pointing at the fork's check-run view (`Run / CI / Preflight`). If the run can't be found, it reports `action_required` with instructions for the contributor to enable GitHub Actions on their fork.
+- **`update_build_status.yml`** (upstream) — runs on a 15-minute cron, walks open PRs, and `PATCH`es each `Build` check's status/conclusion from the corresponding fork run. This is what eventually flips the upstream check from "queued" to "success/failure".
+
+### Contributor checklist
+
+If the upstream `Build` check fails with **"Workflow run detection failed"**:
+
+1. Ensure GitHub Actions is enabled on your fork (Settings → Actions → "Allow all actions").
+2. Rebase onto the latest upstream `main` and force-push — `notify_test_workflow.yml` looks up the run by `(fork owner, fork repo, branch, head sha)`, and a stale base can prevent the match:
+   ```sh
+   git fetch upstream
+   git rebase upstream/main
+   git push origin <branch> --force
+   ```
+3. If the fork's run is healthy but the upstream check stays "queued" longer than ~15 minutes, the cron job in `update_build_status.yml` may have skipped a cycle — pushing a follow-up commit re-triggers `notify_test_workflow.yml` immediately.
+
 ## How to format `.md` document
 
 We are using `prettier` to format `.md` files.