Changelog

All notable changes to Generous.nz are recorded here, grouped by version. Every deployed build has a version number, a git tag, and an entry below.

Versioning scheme

Versions follow MAJOR.MINOR.PATCH:

0.x — pre-launch. Minor bumps mark a completed build phase or major feature set; patches are bug fixes on the current build.
1.0.0 — first public launch.
1.x+ — post-launch: minor = new features, patch = fixes, major = breaking changes or significant re-architecture.

Each version is tagged in git (v0.2.0) and the current version is kept in package.json. See docs/version-control.md for how to cut a release.

Business page copy: "That's $50 in platform fees free from us — a real head start..." now reads "That's $50 in fees we don't charge, giving you a head start while fundraisers discover your products."

Shielded.co.nz site-safety badge added to the footer (with the embed script and modal opener), sitting alongside the AdvertiseOnline.co.nz credit.

Fundraisers can now download a QR code for their public page from the "Edit fundraiser" screen. The QR includes the site's G icon in the centre (white-backed so it stays scannable), links to /f/{slug}, and is served by a new owner/staff-only API with a one-click PNG download.

The audit archive job now verifies the Dedirock copy by listing the remote audit/ directory (a single-file ls didn't match reliably), so the upload-then-delete pipeline completes.

The audit archive job connects to the Dedirock backup server as the gnzbackup SFTP user (matching the live offsite backup), and creates the remote audit/ directory in a tolerated pre-step so an "already exists" reply never aborts the upload.

Audit-log archive and retention job (scripts/audit-archive.sh + audit-archive.mjs): logs older than 30 days are exported to a gzipped TXT archive, pushed to the Dedirock off-site backup server over the restricted SFTP channel, verified, and only then deleted from the database. Scheduled daily so the live audit log keeps a rolling 30-day window.

Blog full-text search now passes the language config as a fixed SQL literal (the previous build parameterized it, which Postgres rejected with "function plainto_tsquery(text, text) does not exist").

Blog search now uses Postgres full-text search (tsvector column + GIN index via migration 0057) ranked by relevance, replacing slow ILIKE text scans over title/excerpt/content — it stays fast as the blog grows past a few thousand posts. English stemming means "voucher" and "vouchers" now match each other.
The public blog API is paginated (page, pageSize up to 50, q) and returns posts, total, page and pageCount instead of every post at once.