Skip to content

docs: preserve 31 untracked blog drafts - #11328

Closed
localai-bot wants to merge 1 commit into
masterfrom
docs/preserve-blog-drafts
Closed

docs: preserve 31 untracked blog drafts#11328
localai-bot wants to merge 1 commit into
masterfrom
docs/preserve-blog-drafts

Conversation

@localai-bot

Copy link
Copy Markdown
Collaborator

Puts 31 blog posts (plus the section _index.md) into git. They span 2024-04-24 to 2026-07-12 and have been sitting untracked in docs/content/blog/ in a local working tree, never committed to any branch.

This changes nothing that is published. All of them return 404 on localai.io today, and the branch is not merged.

Why now

During the session that produced #11324 and #11325, a rebase running in that checkout briefly took the whole directory out of the working tree. It came back when the rebase finished, but that is exactly the failure mode untracked files have, and there was no copy in git to fall back on.

They are clean

A no-ai-slop detect pass over all 31 returns zero hits on every pattern. Not a surprise: they are short release notes, median 83 words, in the same register as what-landed-in-localai-4-8.md, which also needed no edits in #11324. The longest one opens

You record an hour of standup, run it through Whisper, and get back one long wall of text. Every word is correct. You still have no idea who said any of them [...] I've done this more times than I want to admit.

and uses Enough chitchat, let's look at what's in it. That is the real voice, so nothing here was rewritten. Content is byte-identical to the working-tree originals (diff -r verified).

Not resolved, needs your call before merge

  1. Front matter mismatch. These use TOML +++ with an explicit url = "/blog/<slug>/". The five live posts in website/content/blog/ use YAML --- with no url override. These look like leftovers from before the site split in 94d5aff, when the whole site lived under docs/ and the blog was at /blog/.

  2. Merging this would feed them into the docs site. docs/hugo.toml mounts content wholesale, so 31 new pages would enter the docs build. I could not verify the resulting URLs — the docs theme module is not vendored, so a bare worktree cannot build that site (unknown output format "search" for kind "home"). Worth checking before merge, because the url override interacts with baseURL = 'https://localai.io/docs/'.

  3. Where they should live. If they are meant to be published they probably belong in website/content/blog/ with the front matter converted. If they are not, they want draft = true.

I left all three open rather than guess. If you tell me which you want, the conversion is mechanical.

These 32 posts have been sitting untracked in docs/content/blog/ in a local
working tree, never committed to any branch. They are 404 on the live site,
so nothing here changes what is published.

Committing them so they exist in git. During this session a rebase running in
that checkout briefly took the directory out of the working tree, which is
exactly the failure mode untracked files have.

They are clean writing. A no-ai-slop detect pass over all 32 returns zero hits
on every pattern, which is unsurprising since they are short release notes,
median 83 words, in the same register as what-landed-in-localai-4-8.md.

Not resolved here, needs a decision before this is merged:

- They use TOML +++ front matter with an explicit url = "/blog/<slug>/", while
  the five live posts in website/content/blog/ use YAML --- and no url
  override. These look like leftovers from before the site split in 94d5aff,
  when the whole site was under docs/ and the blog lived at /blog/.
- docs/hugo.toml mounts content wholesale, so merging this to master would
  feed 32 new pages into the docs site build. I could not verify the resulting
  URLs because the docs theme module is not available in a bare worktree.
- If they are meant to be published they probably belong in
  website/content/blog/ with the front matter converted. If they are not, they
  want draft = true.

I have left the content byte-identical to the working-tree originals rather
than guess at any of that.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
@mudler mudler closed this Aug 3, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants