This is an automated email from the ASF dual-hosted git repository.
hello-stephen pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/doris-website.git
The following commit(s) were added to refs/heads/master by this push:
new 7c2a694754c [chore](ci) refresh the docs last-update map on every cron
deploy (#4080)
7c2a694754c is described below
commit 7c2a694754c0a552801ddb2e3f85f501ea088150
Author: Mingyu Chen (Rayner) <[email protected]>
AuthorDate: Fri Aug 21 21:15:40 2026 +0800
[chore](ci) refresh the docs last-update map on every cron deploy (#4080)
## Problem
`scripts/last-update/generate.js` needs full git history — in a shallow
clone every file collapses to the HEAD commit, so it exits 1 rather than
write a misleading map.
Cron Deploy checks out shallow, so it could never run the generator. The
published "last updated" footers were whatever the last manual **Refresh
Docs Last-Update Map** PR had committed into
`scripts/last-update/data.json`, and drifted further behind `master`
until someone remembered to trigger that workflow again.
## Change
`.github/workflows/cron-deploy-website.yml`:
- Checkout now fetches full history with `filter: blob:none`.
`actions/checkout@master` is pinned to a **July 2020** commit
(v2.3.1-era) that predates the `filter` input, hence the bump to `v4`.
- A new step regenerates `data.json` before the build. The build is the
only consumer — nothing is committed, so the checked-in map stays the
local-build copy and the deploy's fallback.
## Cost
Measured against `apache/doris-website`, both over the same link:
| | time | size |
|---|---|---|
| `--filter=blob:none` full history (**added** by this PR) | 12.5s | 21
MB |
| `--depth=1` (what the workflow **already** downloads) | 5m40s | 449 MB
|
The expensive part is the HEAD blobs — images and static assets — and
this PR does not touch that path. Full commit history is only ~5% more
bytes on top, and a GitHub runner is far better connected than the
machine those numbers came from. The generator itself runs in ~1s and
uses only Node built-ins, so it needs no `yarn install`.
## Failure behavior
The refresh step is deliberately non-fatal. `generate.js` writes
`data.json` only after computing everything, so a failure leaves the
committed map intact and the deploy still ships with the previous
timestamps. A cosmetic footer should not be able to block a release; the
step emits a `::warning::` instead.
## Notes
- **Refresh Docs Last-Update Map** is kept. The published site no longer
depends on it, but it keeps the committed fallback — what local builds
read — from drifting far behind. Comments in both workflows were updated
to say so.
- `manual-deploy-website.yml` is **not** changed. It still checks out
shallow and will deploy the committed timestamps, which is exactly what
it does today, so this is not a regression. Happy to apply the same two
changes there if reviewers want them consistent.
- The other edits are comment-only, correcting three places that claimed
the deploy renders these dates "without a full-history clone".
🤖 Generated with [Claude Code](https://claude.com/claude-code)
https://claude.ai/code/session_013F7zebd8iJKbAJB2RSB8iq
Co-authored-by: morningman <[email protected]>
Co-authored-by: Claude Opus 5 (1M context) <[email protected]>
---
.github/workflows/cron-deploy-website.yml | 20 +++++++++++++++++++-
.github/workflows/refresh-docs-last-update.yml | 9 ++++++---
docusaurus.config.js | 11 ++++++-----
scripts/last-update/generate.js | 11 +++++++----
4 files changed, 38 insertions(+), 13 deletions(-)
diff --git a/.github/workflows/cron-deploy-website.yml
b/.github/workflows/cron-deploy-website.yml
index 1e0b1185d5b..3ba9185ee0b 100644
--- a/.github/workflows/cron-deploy-website.yml
+++ b/.github/workflows/cron-deploy-website.yml
@@ -10,10 +10,16 @@ jobs:
environment: Production
steps:
- name: Checkout
- uses: actions/checkout@master
+ # v4 (the pinned @master commit predates the `filter` input) with
+ # the full history the last-update map is computed from. Blobless
+ # keeps that affordable: only commit/tree metadata is fetched
+ # (~15MB), not every historical file version (~1.7GB).
+ uses: actions/checkout@v4
with:
persist-credentials: false
submodules: recursive
+ fetch-depth: 0
+ filter: blob:none
- name: Add Chinese Fonts
run: |
@@ -32,6 +38,18 @@ jobs:
with:
node-version: 20
+ - name: Refresh docs last-update map
+ # Recompute scripts/last-update/data.json from the history
fetched
+ # above, overwriting the committed copy so every deploy renders
+ # current "last updated" footers instead of waiting for the
manual
+ # "Refresh Docs Last-Update Map" PR. Only the build below reads
it;
+ # nothing is committed. Non-fatal on purpose - the generator
writes
+ # the file only after it succeeds, so a failure leaves the
committed
+ # map untouched and the site still ships.
+ run: |
+ node scripts/last-update/generate.js \
+ || echo "::warning::last-update refresh failed; deploying
with the committed map"
+
- name: Add Swap Space
run: |
sudo fallocate -l 8G /mnt/swapfile2
diff --git a/.github/workflows/refresh-docs-last-update.yml
b/.github/workflows/refresh-docs-last-update.yml
index cf5c559cd4f..3b0d37889b3 100644
--- a/.github/workflows/refresh-docs-last-update.yml
+++ b/.github/workflows/refresh-docs-last-update.yml
@@ -1,9 +1,12 @@
name: Refresh Docs Last-Update Map
# Manually triggered. Recomputes each document's last-update time from the full
-# git history and opens a PR updating scripts/last-update/data.json. The
regular
-# Cron Deploy then serves those timestamps (docs/dev + 4.x + community footers)
-# without needing full history itself.
+# git history and opens a PR updating scripts/last-update/data.json.
+#
+# The Cron Deploy regenerates that map on every run, so the published site does
+# not depend on this workflow. It exists to keep the committed copy - what
local
+# builds read, and what a deploy falls back to if its own refresh fails - from
+# drifting far behind master.
#
# Requires the repo/org setting "Allow GitHub Actions to create and approve
pull
# requests" to be enabled so the default token can open the PR.
diff --git a/docusaurus.config.js b/docusaurus.config.js
index c4a7cfd3f06..f8361cbda32 100644
--- a/docusaurus.config.js
+++ b/docusaurus.config.js
@@ -61,11 +61,12 @@ const COURSE_RAIL_BOOTSTRAP = `(function () {
}());`;
// Per-document last-update timestamps, generated from git history by
-// scripts/last-update/generate.js and refreshed on demand via the
-// "Refresh Docs Last-Update Map" workflow. markdown.parseFrontMatter (below)
-// injects these as `last_update` front matter so the 2-hourly deploy renders
-// "last updated" without a full-history clone. Missing file /
not-yet-generated
-// map → Docusaurus falls back to its own git lookup.
+// scripts/last-update/generate.js. The 2-hourly Cron Deploy regenerates the
map
+// before it builds, so published dates are always current; the committed copy
+// is what local builds use, and what a deploy falls back to if that refresh
+// fails. markdown.parseFrontMatter (below) injects the dates as `last_update`
+// front matter, sparing the build a per-file git lookup. Missing file /
+// not-yet-generated map → Docusaurus falls back to its own git lookup.
let DOCS_LAST_UPDATE = {};
try {
DOCS_LAST_UPDATE = require('./scripts/last-update/data.json');
diff --git a/scripts/last-update/generate.js b/scripts/last-update/generate.js
index b4db07e5bda..82d66dfd9ea 100644
--- a/scripts/last-update/generate.js
+++ b/scripts/last-update/generate.js
@@ -5,10 +5,13 @@
// derived from this repo's git history.
//
// docusaurus.config.js reads this map at build time
(markdown.parseFrontMatter)
-// and injects it as each document's `last_update` front matter, so the regular
-// 2-hourly deploy can render "last updated" without a full-history clone or a
-// per-file git lookup. Refreshed on demand by the "Refresh Docs Last-Update
-// Map" workflow (.github/workflows/refresh-docs-last-update.yml).
+// and injects it as each document's `last_update` front matter, so the build
+// never does a per-file git lookup. The 2-hourly Cron Deploy runs this script
+// itself (its checkout uses fetch-depth: 0 + filter: blob:none) and builds
from
+// the result, so published dates track master. The committed data.json is the
+// local-build copy and the deploy's fallback; the "Refresh Docs Last-Update
+// Map" workflow (.github/workflows/refresh-docs-last-update.yml) keeps it
fresh
+// via PR.
//
// Requires full git history to be accurate. In a shallow clone every file
// resolves to the single HEAD commit, so run this only where the history is
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
