Ministra stops being a module — the whole Stalker portal (portal.php,
MinistraBootstrap, PortalHandler/PortalHelpers and the STB front-end) now
lives in src/Ministra/ under the XcVm\Ministra namespace, served at
/home/xc_vm/Ministra via the nginx alias.
- src/ministra/* and Modules/ministra_85a7d/{PortalHandler,PortalHelpers}
→ src/Ministra/; MinistraModule.php + module.json removed. Ministra was
the only committed module, so src/Modules/ keeps a .gitkeep.
- portal.php resolves PortalHandler as a sibling and derives MAIN_HOME from
its new location (glob crutch gone).
- nginx alias + AuthRepository $rAlias switched to /home/xc_vm/Ministra
(PascalCase); ministra entry dropped from bundled_modules.php.
- Makefile: Modules/ removed from LB_DIRS — all modules are MAIN-only, so
the ~50 MB of portal assets no longer ship to LB nodes.
- ArchitectureTest: zero committed modules is now a valid state.
- PHPStan: analyse src/Ministra, exclude the procedural portal.php entry,
repath the ministra baseline entries.
- Docs (architecture, ministra-browser-emulation, extraction plan) updated
to the new layout; the "extract to a separate repo" plan is cancelled.
Verified: php -l, make gates, make phpstan (No errors), full unit suite
(432 tests). On-server smoke: handshake + get_profile work end-to-end with
a registered MAC after deploy.
BoundaryInterface was a marker interface with no runtime consumer —
nothing read getEntryPoint()/isIsolated() and, being static-less
metadata, it enforced nothing. Its only implementor was MinistraModule.
Removes the interface, drops `implements BoundaryInterface` plus the two
orphaned methods from MinistraModule (getName/getVersion stay — they come
from BaseModule/ModuleInterface), and deletes the now-empty Core/Boundary/.
Test contract (InterfaceContractTest) loses the three BoundaryInterface
assertions; docs (en/ru architecture + modules, .github instructions) now
describe isolated subsystems like Ministra as a convention — own entry
point + bootstrap — rather than a marker interface.
Verified: php -l, make gates, make phpstan (No errors);
InterfaceContractTest + ArchitectureTest green (33 tests, 96 assertions).
The panel is deeply coupled to TMDb (VOD import, player metadata, admin
search, two crons), so shipping it as an uninstallable module only added
failure modes: after the move to hash-suffixed dirs (tmdb_f4e6e) every
hardcoded `Modules/tmdb/lib/...` require broke, and 2.3.3 crons died with
"Failed opening required TmdbClient.php".
tmdb -> core:
- Vendored \TMDB client -> src/Infrastructure/Tmdb/lib/; the only loader
is TmdbApiService::requireLibrary() (now public, also loads Release.php).
- TmdbApiService -> XcVm\Infrastructure\Tmdb — composer-autoloaded in every
bootstrap context, no module boot required (player scope never booted
modules, so module-namespace classes were unreachable there).
- TmdbCron / TmdbPopularCron -> XcVm\Domain\Vod; cron jobs -> Cli/CronJobs
(picked up by the console.php scan; command names cron:tmdb and
cron:tmdb_popular are unchanged).
- TmdbController -> Public/Controllers/Admin; tmdb_search / tmdb api
actions registered in routes/admin.php (same dispatchApi fallback).
- Domain/Vod services and player_functions.php load the lib through
TMDbService::requireLibrary() instead of hardcoded module paths.
- tmdb removed from config/bundled_modules.php. ModuleLoader gains
CORE_PROVIDED_MODULES: released watch/plex archives still declare
"dependencies": ["tmdb"] — such deps are stripped during manifest
normalization and in ModuleManager::listModules().
- syncBundledModules() purges stale on-disk tmdb module dirs and their
config/modules.php state on upgraded panels, so the old copy cannot boot
alongside the core implementation and collide on command names.
Standard-set provisioning fix (root cause of the "Undefined variable $db"
errors from watch/plex settings views on 2.3.3):
- Production still ran watch_e6c86/plex_20cd9-less legacy copies migrated
from 2.3.2 with generated hash_ids; provisionStandardSet() treated any
same-name directory as "already on disk" and never fetched the pinned
1.0.2/1.0.1 releases that contain the fix. A same-name directory whose
identity does not match the pinned hash_id is now considered stale: it
is deleted and the pinned release is installed in its place.
- installModuleFromSource(): when the module is already recorded as
installed (files re-provisioned over a stale copy), run updateModule()
(incremental from->to migrations) instead of re-running the initial
install.
- ArchitectureTest/InterfaceContractTest resolve modules via module.json instead
of the directory basename (works with {name}_{hash5}).
- ModuleLoaderTest: loads from a hash-suffixed dir; a broken module and its
dependents are skipped without aborting the whole load.
- ModuleManagerMigrationsTest: legacy bare -> hashed migration (+ stale-dup drop),
and install of a module living in a hash-suffixed directory.
Namespace the modules' own classes into XcVm\Module\<Pascal> (Plex, Watch, Tmdb,
Ministra) — PlexController/PlexService/PlexCron/..., WatchController/WatchService/...,
TmdbController/TmdbCron/TmdbApiService/..., PortalHandler/PortalHelpers (~23 classes).
They now resolve through the Phase-2 ModuleLoader PSR-4 resolver.
- The bundled third-party tmdb/lib/* (TMDB client, Entities, roles, config) stays
GLOBAL, like the other vendored libs; namespaced Tmdb classes reference it via
\TMDB etc.
- Rewrite the *Module classes' (and any sibling) 'use ShortName;' imports → FQCN;
add use to referrers; convert leading-backslash refs. Sub-classes keep the
Core/Domain/Cli use imports added during those layers.
- Qualify built-ins (\DateTime, \Exception, \PDO, ...) and the global lib classes.
- phpstan-baseline.neon regenerated.
Verified: php -l clean; PHPStan no errors; PHPUnit 295/295; all 4 modules load and
their sub-classes (e.g. XcVm\Module\Plex\PlexService) resolve via the module
PSR-4 autoloader; re-sweep clean.
Divarion_D