95135[TASK] Add autocomplete to setup and language commands
Commit message
Review
-
Code review CR +2 Done
- Georg Ringer +2
- Stefan Bürk +1
- Sebastian Fischer +1
- Benni Mack +1
-
Verified V +2 Done
- Georg Ringer +2
- Stefan Bürk +1
- Benni Mack +1
- core-ci +1 CI
Issues 1
Files 2
Review log 1316
-
Georg Ringer uploaded patch set 1 08/06/26, 19:35
-
Georg Ringer commented on patch set 1 08/06/26, 19:35
This change is ready for review. -
Georg Ringer updated the commit message in patch set 2 08/06/26, 19:37
-
core-ci voted on patch set 1 V +1 08/06/26, 19:45 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/106031
-
core-ci voted on patch set 2 V +1 08/06/26, 19:47 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/106032
-
Tobias Gaertner voted on patch set 2 CR +1 V +1 08/06/26, 20:02
-
Simon Schaufelberger voted on patch set 2 CR +1 V +1 08/15/26, 09:52
-
Stefan Bürk commented on patch set 2 CR -1 08/15/26, 18:07
Thanks for pushing the completion series forward - the declarative $suggestedValues closure is the right tool and --server-type is exactly how this should look. Three things before this can go in. 1. "Resolve:" must be "Resolves:". Build/git-hooks/commit-msg requires ^Resolves: #[0-9]+$, and the typo already had a visible effect: #107249 is still in status New with no Gerrit Code Review journal entries at all, whereas #107247 (95143/95147, with the correct keyword) is Under Review with the usual automatic push notes. On merge this would not close the issue. While touching the message: it announces "language:update: locales" and "setup: driver", but the patch also adds completion for setup --distribution and setup --server-type. And a sharper subject would help - "[TASK] Add autocomplete to commands of EXT:install" (49 chars) mirrors 95143 and the forge issue title. 2. --distribution suggests values the command rejects. The closure returns array_keys($distributions['active'] + $distributions['inactive']), but getDistributionForActivation() validates against inactive only and throws 'Distribution "%s" is not available.' for everything else - and it is only reached when $distributions['active'] === [], so an active key can never be a usable value. It also ignores the impexp guard that the option's own description already carries ("[disabled] Requires typo3/cms-impexp to be installed"); with impexp inactive, any value throws 1775034287. Suggest returning [] when the impexp package is not active, and otherwise only array_keys($this->setupService->getAvailableDistributions()['inactive']). 3. language:update should not call loadExtLocalconfDatabase(). By the time a suggestion closure runs, CommandApplication::run() has already done the full boot - _complete is not in CommandRegistry, so wantsFullBoot() is true. You can see it on the already-merged impexp:export completion, which reads TcaSchemaFactory and works: "php bin/typo3 _complete --no-interaction -sbash -c2 -a1 --input=bin/typo3 --input=impexp:export --input=--table=" returns the full table list here. Calling it a second time re-requires the cached ext_localconf (plain require, see SimpleFileBackend::require()) and purges all singletons. And it is not needed anyway: LanguagePackService::getActiveLanguages() is a bare $GLOBALS['TYPO3_CONF_VARS']['LANG']['availableLocales'] read. The bit that actually worries me is the uninstalled system: run() deliberately skips the boot when Bootstrap::checkIfEssentialConfigurationExists() is false, and the closure would then be the only caller of a method whose docblock says "can potentially fatal ... Use only if really needed" - from a TAB press. Please use the 92733 form and just resolve LanguagePackService directly. The same remark applies to 95143's getAvailableCacheGroups(), though there the boot is at least semantically needed for extension-registered cache groups. Two smaller points, not blocking: * --driver offers all six connectionLabels keys, but the validator checks $driverOptions[... . 'ManualConfigurationOptions'], which SetupDatabaseService fills only behind DatabaseCheck::isMysqli()/isPdoMysql()/isPdoPgsql()/isPdoSqlite(). On a pdo_mysql-only container, "--driver=<TAB>" still offers postgres. The interactive ChoiceQuestion has the same flaw today, so this is pre-existing - but you are one line away from fixing both. * UpdateLanguagePackCommand is EXT:core, while #107249 is scoped to EXT:install. Either move that hunk to 95143/#107247 or widen the issue, otherwise the per-extension split of the series does not hold. Finally: --driver and --server-type are testable with Symfony\Component\Console\Tester\CommandCompletionTester without any boot, and core has no completion test yet (grep -rn CommandCompletionTester typo3/ is empty) even though the #106839 changelog links Symfony's testing docs. A small typo3/sysext/install/Tests/Unit/Command/SetupCommandTest.php here would establish the pattern for the rest of the series - and would have caught point 2. -
Sebastian Fischer voted on patch set 2 CR +1 08/20/26, 07:11
-
Georg Ringer uploaded patch set 3
CR +1V +108/27/26, 16:09 -
core-ci voted on patch set 3 V +1 08/27/26, 16:17 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/107280
-
Sebastian Fischer voted on patch set 3 CR +1 08/31/26, 14:05
-
Stefan Bürk commented on patch set 3 CR +1 V +1 09/02/26, 12:14
Reply
Done -
Benni Mack voted on patch set 3 CR +1 V +1 09/02/26, 12:16
-
Georg Ringer voted on patch set 3 CR +2 V +2 09/02/26, 13:02
-
Georg Ringer merged patch set 4 09/02/26, 13:49