Skip to content

Release 1.19.0 - #171

Merged
turegjorup merged 43 commits into
mainfrom
release/1.19.0
Oct 26, 2026
Merged

turegjorup merged 43 commits into
mainfrom
release/1.19.0

Conversation

@turegjorup

Copy link
Copy Markdown
Contributor

Release 1.19.0

  • #170 Update the docker compose files from the itkdev-docker templates, which use the maintained nginx/nginx-unprivileged image
  • #169 Mark advisories as not affected per git repo, and sort sites by their open advisories
  • #168 Rewrite the fixtures to cover all content with made-up data shaped like production
  • #167 Let the latest dir result decide which sites exist, and purge detection results not seen for 30 days every night
  • #166 Update EasyAdmin to 5.7
  • #165 Update Composer packages, including symfony/expression-language for CVE-2026-109449
  • #164 Clear two Doctrine deprecations

Minor: advisory assessments (a new feature), and the dir cleanup and nightly purge.

Deploy notes:

  • One migration: the advisory_assessment table.
  • The workers consume scheduler_default instead of scheduler_drupal_releases.
  • The nginx image changes to nginx/nginx-unprivileged:alpine, so pull images on deploy.
  • After deploy, wait for one harvest, then run bin/console app:replay:detection-results --type docker to restore the 27 sites deleted by the 2026-10-19 dir replay.

A string direction is deprecated in doctrine/orm 3.
MariaDB commits implicitly on every schema change, so the wrapping transaction was already committed and doctrine/migrations deprecates silencing that.
fix: clear Doctrine sort direction and migration transaction deprecations
Ignore dir results older than the server's latest, so replays cannot apply an old list. Attach sites to the listed dir they are in instead of a LIKE prefix, where _ was a wildcard, and delete sites and results outside every listed dir.
An unchanged dir result only had its timestamp updated, so sites recreated after it was first processed, e.g. by a docker replay, were never removed.
Run app:data:purge on the default schedule at 03:00 with up to 30 minutes jitter. --date now takes relative dates such as "30 days ago". The worker consumes scheduler_default.
Attribute tasks such as the purge cannot be stateful on their own, and the worker restarts every 15 minutes, so the jittered purge would skip most days. Turn DrupalReleaseSchedule into the stateful default schedule that attribute tasks join, and consume only scheduler_default.
The worker resets services after each message, which replaces a closed EntityManager, and failed messages are retried and then sent to the failed transport. Stopping on every failure only restarted the worker.
8669: Clean up removed sites and purge old detection results
The site filter test now removes the fixture sites it does not set up, the link command test counts from what the fixtures already link, the module sort test checks the order of the counts, and the linker test checks the module is not linked by name rather than not linked at all.
Made-up data shaped like production: a repo checked out at two versions with overlapping advisories, a repo without advisories, a release deployment without a repo, drupal.org and Packagist advisories with and without a CVE, one CVE under two advisory ids, and Drupal modules linked to their Composer packages. A test pins these cases and that only example data is used.
One file per area: servers, sites with the detection results behind them, git repos, packages and advisories, Docker images, projects with code owners and security contracts, OIDC and service certificates. Every entity, detection type, server type and site type is represented, with made-up values shaped like production.

FixturesTest pins the scenarios, that replaying the detection results rebuilds the same installations and sites, that each server's dir result lists all its installations, and that only example data is used. Tests tied to the old fixtures now use the shop sites, the never-harvested gpu-01 server or counts relative to the fixtures.
On a site's detail page, an advisory can be marked as not affected with a reason, one at a time or all current ones at once, and a mark can be removed. A mark belongs to the git repo, so it covers every checkout of it, and sites without a repo cannot be marked.

Marked advisories no longer count on the site list, in its sort or on the repo advisories page. Marks are stored by repo and advisory name, so truncating derived data and replays leave them in place.
@turegjorup turegjorup self-assigned this Oct 26, 2026
@github-actions

Copy link
Copy Markdown

API Specification - Non-breaking changes

No changelog changes

@turegjorup
turegjorup merged commit 2c51fdf into main Oct 26, 2026
22 checks passed
@turegjorup
turegjorup deleted the release/1.19.0 branch October 26, 2026 15:40
@codecov-commenter

Copy link
Copy Markdown

Codecov Report

❌ Patch coverage is 97.75281% with 4 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (main@7606a43). Learn more about missing BASE report.

Files with missing lines Patch % Lines
src/Entity/AdvisoryAssessment.php 92.85% 2 Missing ⚠️
src/Admin/SemverSort.php 50.00% 1 Missing ⚠️
src/Command/ReplayDetectionResultsCommand.php 0.00% 1 Missing ⚠️
Additional details and impacted files
@@           Coverage Diff           @@
##             main     #171   +/-   ##
=======================================
  Coverage        ?   76.77%           
  Complexity      ?     1346           
=======================================
  Files           ?      172           
  Lines           ?     4121           
  Branches        ?        0           
=======================================
  Hits            ?     3164           
  Misses          ?      957           
  Partials        ?        0           
Flag Coverage Δ
unittests 76.77% <97.75%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
  • 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

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