4 Commits
Author SHA1 Message Date
PaddyandClaude Opus 5.5 b22e89001a Run tests and deploy in one Gitea Actions job
Test and deploy / test-and-deploy (pull_request) Successful in 2m25s
Two jobs meant checking out, setting up Node and running npm ci (and in the
frontend the build) twice, about two minutes each, because every job starts
in a fresh container. Now one job tests and, for pushes to master only,
deploys. The push-to-master condition is spelled out on every deploy step
instead of an env flag compared to 'true', which would skip the deploy
silently and stay green if the boolean were rendered differently. A failed
test still stops the deploy (an if: without a status function implies
success()).

Docs: the one-job layout, and that Jenkins deployed as root - files it left
in an app folder must be chowned to the subscription user before the first
Gitea deploy (hit on the API).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-08 22:51:26 +01:00
PaddyandClaude Opus 5.5 38eff139be Document the real deploy setup: SSH port, deploy user, shared key
Test and deploy / test (pull_request) Successful in 2m29s
Test and deploy / deploy (pull_request) Skipped
- SSH on port 42069, deploy user plesk-nachklang, one deploy key for all
  Nachklang repos as a Gitea organisation secret (host, port and user as
  organisation variables; path and build values per repository).
- Key-only login for the deploy user via a Match block in sshd_config.
- Plain /bin/bash instead of Plesk's chrooted shell: bash segfaults in the
  chroot copy on this server, so every session closed right after login;
  the read-access trade-off and a stricter rrsync option are noted.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-08 21:46:54 +01:00
PaddyandClaude Opus 5.5 b0cd657fd9 Deploy with Gitea Actions instead of Jenkins
Tests every push and pull request (lint, typecheck, Vitest, production
build); pushes to master build the standalone bundle in the runner and
rsync it over SSH as the subscription user into the subdomain's Plesk
folder, then restart Passenger via tmp/restart.txt. No mounted host paths;
rsync --delete leaves tmp/ and .well-known/ alone. Optional smoke test once
DEPLOY_SMOKE_URL is set.

Setup (deploy key, pinned host key, Gitea secrets and variables, Plesk
Node.js settings) and the switchover order are in docs/deployment.md.
Checked locally with actionlint and against a throwaway SSH server, which
also caught "~" in the rsync -e string resolving to the wrong home.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-08 21:09:19 +01:00
PaddyandClaude Opus 5.5 0cda40f85e Rewrite the calendar frontend in Next.js
Replaces the Angular 18 app with Next.js 16 / React 19 / Tailwind 4 / shadcn,
the same stack and design system as nachklang-feedback, planned in
docs/rewrite-plan.md.

- Admin: all calendars in one mobile-friendly list (month groups, calendar
  chips, Kommend/Vergangen/Alle, search, sort, filters in the URL), an edit
  sheet for create/edit/duplicate/delete, publish confirmation, and an
  "Abonnieren" dialog with subscription links for every calendar.
- Public landing page: subscribe links from env instead of hardcoded, plus
  the next public events, server-rendered and revalidated every 5 minutes.
- All times handled explicitly as Europe/Berlin; whole-day events stored as
  00:00-23:59. Fixes carried over from the Angular app: the move dialog
  leaking a cancelled calendar change, saves that looked saved when they
  failed, in-place Date mutation, and duplicated calendar-id mappings.
- Same shared-session auth as before (proxy.ts gate, /admin/me, 401 vs 403).
- 112 Vitest tests over lib/; docs in docs/ (architecture, features,
  gotchas, API wish list, deployment).

Needs the API branch feature/calendar-ical-escaping-subscriptions deployed
first (GET /calendar/events/subscriptions).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-28 22:15:59 +02:00