Skip to content

Release 1.1.1 - #45

Merged
jalmena merged 7 commits into
mainfrom
release/1.1.1
Oct 5, 2026
Merged

jalmena merged 7 commits into
mainfrom
release/1.1.1

Conversation

@jalmena

@jalmena jalmena commented Oct 5, 2026

Copy link
Copy Markdown
Owner

Release 1.1.1

The app updates itself: the service worker takes over as soon as a new version is installed and reloads the open pages, so an updated server no longer looks like the version before it until every page is closed. This is what kept 1.1.0's changes (the marks' trash action, the body-map selection fixes, "vistazo", the design pass) out of sight on an upgraded installation. Continuous integration now runs on GitHub-hosted runners in two stages, about ten minutes a run.

The changelog is in CHANGELOG.md. Merged locally with a signed-off merge commit once the checks pass; the release workflow publishes ghcr.io/jalmena/nevus:1.1.1, tags the merge commit and creates the GitHub release.

Signed-off-by: Jose David <josedalmena@gmail.com>
…inds open

The service worker waited for a prompt the interface never showed, so
after the server was updated the old worker kept serving its cached
shell until every page of the app had been closed: 1.1.0 looked like
the version before it. The new worker now takes over as soon as it is
installed, claims the open pages and, when it is an update rather than
the first install, reloads them so they run the new shell. The offline
outbox lives in IndexedDB and survives the reload.

Signed-off-by: Jose David <josedalmena@gmail.com>
The repository is public, so GitHub's runners cost nothing and many run
in parallel; a run that took an hour on the two-core self-hosted runner
takes about ten minutes. The cheap checks go first (commit messages,
secrets and dependency audits, the frontend's lint, types, tests and
build) and the three expensive jobs run side by side only when they
pass, so a failure in the first stage stops there. The browser used by
the end-to-end job is cached between runs. The self-hosted runner stays
registered and the workflow keeps its tolerances, so it can take the
work again by changing runs-on.

Signed-off-by: Jose David <josedalmena@gmail.com>
Signed-off-by: Jose David <josedalmena@gmail.com>
The verification job stamps the real clock, while the test simulated
days from a fixed date in early October; once the calendar passed that
date, "eight days later" was less than a week after the real stamp and
the test failed. The simulated days now start tomorrow.

Signed-off-by: Jose David <josedalmena@gmail.com>
Signed-off-by: Jose David <josedalmena@gmail.com>
Signed-off-by: Jose David <josedalmena@gmail.com>
@jalmena
jalmena merged commit 38c4662 into main Oct 5, 2026
6 checks passed
@jalmena
jalmena deleted the release/1.1.1 branch October 5, 2026 15:29
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.

1 participant