Skip to content

roadmap.md

Latest commit

 

History

History
51 lines (35 loc) · 2.3 KB

roadmap.md

File metadata and controls

51 lines (35 loc) · 2.3 KB
Raw

Apache File Directory — Roadmap

Known gaps, observed from the repository. Limitations, not a schedule.

Not gaps

Two things look like omissions and are not:

It does not generate listings. Being a hand-written static page is the entire design. Adding generation would need a server, which would make it the thing it is imitating rather than a copy of it.

The markup is not modern. HTML 3.2, table layout, valign attributes. Modernising it would produce a page that no longer looks like Apache output, which is its only purpose.

Both are recorded here so nobody "fixes" them.

Real gaps

Nothing validates the listing. Links, dates, and sizes are all hand-typed, and no check exists for a row pointing at a file that is not there. A small script comparing rows against a directory would catch the most common mistake without changing the static design.

The example rows are a trap. The shipped index.html includes samples, and forgetting to delete them leaves a listing advertising files that do not exist. A commented-out block rather than live rows would make the mistake impossible.

icons/ is easy to leave behind. Copying index.html alone breaks every image. Inlining the icons as data URIs would make the page genuinely single-file, at the cost of readability and of matching what Apache emits.

No tests, and none is obviously right. Defensible for one static file, but there is currently nothing that would catch a malformed table or a broken doctype.

Repository hygiene

Per-repo issue templates override the org defaults. .github/ISSUE_TEMPLATE/ still holds Markdown templates from before the org-level .github repository supplied YAML forms. GitHub prefers the local ones, so this repository does not get the shared set. Removing them would inherit the org versions.

The workflow filename differs from its siblings. Worth a glance when comparing the three file-directory repositories — see Chrome and Nginx.

Non-goals

  • Being a real directory index. Apache already does that.
  • JavaScript. No behaviour is wanted; adding any would break running from file://.
  • A build step. The value of the project is that deployment is copying two paths.