commits
Stores as a cookie and in local storage as a fallback. Passes from origin to sequoia.pub.
Co-Authored-By: @stevedylan.dev
Also added wrangler to the devDependencies so you can just run `bun run dev:api` and everything just works.
Also fixes a bug I just noticed in the Subscribe help page.
`updateDocument` previously built a fresh record from frontmatter only and called `putRecord`, silently dropping any PDS-side fields such as `bskyPostRef`. This broke the `<sequoia-comments>` web component after the first re-publish.
Fix: fetch the existing record with `getRecord` before writing, spread its fields into the new record, then overwrite with the fresh frontmatter-derived values. This preserves all PDS-side fields while still updating the document content correctly.
Fixes stevedylandev/sequoia#5
Biome can use .editorconfig, so use consistent settings. Also refactors biome.json to use ubiquitously in the project.
Unfortunately, markdown files - which we can hopefully associate .mdx files with - aren't supported yet. See https://github.com/biomejs/biome/issues/3718.
Fixes #15, parsing YAML and TOML using dedicated libraries that support multiline strings correctly.
Also adds tests that were run before and after changes. Multiline string tests expectedly failed before.
Added a `--verbose` flag to `publish` that shows the post URL. Unfortunately, the resolved source path may not be the rended page path, but at least the CLI will show it now.
- Adds interactive command to update either the sequoia.json config or
the atproto publication record
Adds a new `stripDatePrefix` configuration option that removes
YYYY-MM-DD- prefixes from filenames when deriving slugs. This
supports Jekyll-style post naming where files like
`2024-01-15-my-post.md` should produce slug `my-post`.
`updateDocument` previously built a fresh record from frontmatter only and called `putRecord`, silently dropping any PDS-side fields such as `bskyPostRef`. This broke the `<sequoia-comments>` web component after the first re-publish.
Fix: fetch the existing record with `getRecord` before writing, spread its fields into the new record, then overwrite with the fresh frontmatter-derived values. This preserves all PDS-side fields while still updating the document content correctly.
Fixes stevedylandev/sequoia#5