Moduletrace/bunext
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Date sitemap lastmod from the build pipeline; warn on skipped pages

Browse Source 
- bunext build records when each page's client bundle hash last changed
  in .bunext/page-lastmod.json (sitemap.lastmod_file to relocate). The
  hash covers the page's whole import graph, so component edits re-date
  every page that uses them. The sitemap uses these dates as lastmod
- Add sitemap.default_lastmod
- BREAKING: use_file_lastmod now defaults to false (file mtimes are reset
  by checkouts); it is a fallback after the build record
- Warn when a page or __root.tsx is skipped because its default export
  is not at the start of a line (e.g. export { Page as default })
- Replace leaking module mocks in tests (log, grab-root-file-path) with
  spies and fixtures

Co-Authored-By: Claude Opus 5.5 <[email protected]>
This commit is contained in:
tben
2026-10-14 18:25:38 +01:00
co-authored by Claude Opus 5.5
parent be4bebaa22
commit 3a75880ae4
26 changed files with 877 additions and 54 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+21 -1
View File
@@ -39,6 +39,7 @@ The goal is a framework that is:
- [Sitemap](#sitemap)
- [Page `sitemap` Export](#page-sitemap-export)
- [`bunext sitemap` Command](#bunext-sitemap-command)
- [Last-Modified Dates](#last-modified-dates)
- [Large Sitemaps](#large-sitemaps)
- [Generating at Runtime](#generating-at-runtime)
- [robots.txt](#robotstxt)
@@ -702,6 +703,23 @@ bunext sitemap --origin https://staging.example.com
bunext sitemap --no-robots # sitemap only
```
#### Last-Modified Dates
`lastmod` comes from the build pipeline. On every `bunext build`, Bunext compares each page's client bundle hash with the previous build and records the build date for pages whose bundle changed in `.bunext/page-lastmod.json`. The hash covers the page's whole import graph, so editing a shared component re-dates every page that uses it, while untouched pages keep their date. Editing `__root.tsx` changes every page.
Each URL's `lastmod` is the first of:
1. The entry's own `lastmod` (from `additional_paths` or a page's `sitemap` export)
2. The page's date from the build record (also used for entries returned by that page's `sitemap` export)
3. The page file's modification time, if `use_file_lastmod: true`
4. `default_lastmod`
With no date, `<lastmod>` is left out. Run `bunext build` before `bunext sitemap`; the command warns when there is no record.
- **The record must persist between builds.** Without the previous record, every page is dated with the current build. Building on the server you deploy to keeps `.bunext/` between builds. For CI builds from a fresh checkout, set `sitemap.lastmod_file` to a path you cache between runs or commit (it is sorted JSON for small diffs).
- **Server data is not covered.** Content loaded in a server function (e.g. from a database) does not change the bundle, so set `lastmod` on those entries in the page's `sitemap` export.
- Upgrading esbuild or changing build options can change every hash once, re-dating all pages.
#### Large Sitemaps
A sitemap file may hold at most 50,000 URLs and 50MB. When the sitemap has more than `chunk_size` URLs (default 10,000), `bunext sitemap` splits it into chunk files and makes `sitemap.xml` a sitemap index pointing to them:
@@ -756,7 +774,9 @@ export default async function handler(req: Request) {
| `origin` | `string` | Base URL for each `<loc>`. Defaults to `config.origin`, then `http://localhost:<port>`. |
| `exclude` | `(string \| RegExp)[]` | Exact paths, subtrees ending in `/*` (`/admin/*` also excludes `/admin`), or regexes tested on the path. |
| `additional_paths` | `entries \| () => entries \| Promise<entries>` | Extra paths or entries (`path`, `lastmod`, `changefreq`, `priority`), or a sync/async function returning them, e.g. for dynamic routes. Each entry is a `string` or `BunextSitemapEntry`. |
| `use_file_lastmod` | `boolean` | Use each page file's modification time as `lastmod`. Default `true`. |
| `default_lastmod` | `Date \| string` | `lastmod` for entries with no other date. See [Last-Modified Dates](#last-modified-dates). |
| `lastmod_file` | `string` | Page date record written by `bunext build`. Default `.bunext/page-lastmod.json`. |
| `use_file_lastmod` | `boolean` | Fall back to each page file's modification time. Default `false`. |
| `trailing_slash` | `boolean` | Append `/` to every non-root path. Default `false`. |
| `default_changefreq` | `BunextSitemapChangeFreq` | `changefreq` for entries that do not set one. |
| `default_priority` | `number` | `priority` (0.0 to 1.0) for entries that do not set one. |