mirror of
https://github.com/karatektus/pl_mail.git
synced 2026-10-25 01:54:10 +00:00
Paul LütznerandClaude Fable 5
8533993c4b
Settings → Insights lists whatever the extractor registry holds — name, hint and an On/Off segment per extractor, so a new extractor appears here by existing. Off is stored as the disabled set, so next release's extractor starts working without anyone finding a new toggle. The endpoint validates keys against the registry and carries the same CSRF discipline as its neighbours; an extractor without catalogue lines falls back to its key rather than printing a translation path. Handbook gains "The radar" in both languages. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
56 lines
2.9 KiB
YAML
56 lines
2.9 KiB
YAML
# Services that exist only while APP_ENV=test.
|
|
#
|
|
# Imported automatically by MicroKernelTrait (config/services_{env}.yaml) and by
|
|
# nothing else, so none of this is compiled into a dev or prod container. That
|
|
# is the whole reason the file exists rather than a guard inside a class in
|
|
# src/: a stub in the application tree ships, and a stub that ships is a stub
|
|
# somebody eventually reaches by accident.
|
|
#
|
|
# One entry, and it needs the paragraph. Subscribing to a remote calendar cannot
|
|
# be exercised from a browser without a driver a fixture can script — Playwright
|
|
# drives real HTTP, so the collaborators come from the real container, and the
|
|
# object-level FakeCalendarSyncDriver a service test hands in by hand is
|
|
# unreachable from there. See ScriptedCalendarSyncDriver for why pointing a spec
|
|
# at a real (or stubbed) CalDAV server was rejected instead.
|
|
#
|
|
# The priority is load-bearing. Drivers arrive as a tagged iterator and the
|
|
# first that claims a source wins; CalDavCalendarDriver claims every
|
|
# Provider::CalDav connection, so a stub tagged at the default 0 would be asked
|
|
# second and never answer. It claims only connections carrying its own settings
|
|
# key, so being asked first costs the real driver nothing.
|
|
services:
|
|
App\Tests\Support\Calendar\ScriptedCalendarSyncDriver:
|
|
tags:
|
|
- { name: 'app.calendar_sync_driver', priority: 100 }
|
|
|
|
# The same shape, for the same reason, one layer over. DataResetter and both
|
|
# push registries are `final readonly` and cannot be mocked — deliberately —
|
|
# so proving that a reset hands its registrations back BEFORE it truncates
|
|
# the credentials needed to do so means joining the tagged iterator that
|
|
# feeds them.
|
|
#
|
|
# Priority again load-bearing, and again safe: this claims only accounts
|
|
# whose address carries ScriptedPushManager::MARKER, so being asked first
|
|
# costs the real Gmail and Graph managers nothing. A stub that claimed every
|
|
# account would silently take over push for the entire suite.
|
|
App\Tests\Support\Push\ScriptedPushManager:
|
|
public: true
|
|
tags:
|
|
- { name: 'app.push_subscription_manager', priority: 100 }
|
|
|
|
# And the calendar half, for the removal paths that never revoked at all.
|
|
App\Tests\Support\Push\ScriptedCalendarPushManager:
|
|
public: true
|
|
tags:
|
|
- { name: 'app.calendar_push_manager', priority: 100 }
|
|
|
|
# One guaranteed insight extractor, so the settings page and the toggle
|
|
# endpoint can be tested against a registry that is never empty — the
|
|
# production catalogue is free to grow and shrink without those tests
|
|
# noticing. Inert by construction: its supports() never says yes. No
|
|
# priority override needed; it claims no mail, so its place in the order
|
|
# is cosmetic and it sorts itself last.
|
|
App\Tests\Support\Insight\ScriptedInsightExtractor:
|
|
tags:
|
|
- 'app.insight_extractor'
|