Review main Merged

96089[TASK] Add conformance suite for Extbase language overlays

Alexander Schnitzler CR +2 V +2 Hard +469 −0 updated 10/14/26, 16:56

Commit message

Extbase translation behaviour today is documented only by scattered
tests that assert SQL fragments or a single scenario each. Nothing
says, in one place, which records a query returns for a given
language and overlay mode, or what a fetched record's identity
fields look like once translated.

This commit adds Tests/Functional/Persistence/BehaviourConformanceTest.php
with its fixture (Fixtures/BehaviourConformanceTestImport.csv, storage
page 20, five posts and their DA/DE translations). It pins, as outcomes
rather than SQL: which rows a root query returns per overlay mode
(off/on/on-with-floating/mixed) and language, including rows the
overlay drops; the identity fields (uid, _localizedUid, _languageUid)
of a fetched record and of its n:1 and m:n relations, loaded under the
parent's language aspect; 1:n children loaded by the translated parent
uid; and that findByUid() on a default and a translated uid in one
request yields two distinct objects sharing a uid, a known hole with
no prior test. Two known quirks are pinned as they are today and
marked for their fix: count() and LIMIT do not account for rows the
overlay drops afterwards.

This does not change any class under Classes/; it only adds tests.

This is step 1 of 4 of "Pin Extbase persistence behaviour in an
outcome-based conformance suite", which gives the refactorings and bug
fixes that follow an oracle for what Extbase does today in terms of
records and rows, not SQL strings.

Review

  • Code review CR +2 Done
    • Benni Mack +2
    • Georg Ringer +1
    • Simon Schaufelberger +1
  • Verified V +2 Done
    • Benni Mack +2
    • Georg Ringer +1
    • Oliver Klee +1
    • Simon Schaufelberger +1
    • core-ci +1 CI

Issues 3

Files 2

Review log 1924

  1. Alexander Schnitzler uploaded patch set 1 09/25/26, 18:24

  2. Alexander Schnitzler updated the commit message in patch set 2 09/25/26, 18:30

  3. core-ci voted on patch set 1 V -1 09/25/26, 18:33 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/109290

  4. core-ci voted on patch set 2 V -1 09/25/26, 18:40 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/109291

  5. Alexander Schnitzler uploaded patch set 3 V -1 09/25/26, 18:41

  6. Alexander Schnitzler wrote 09/25/26, 18:42

    Topic set to extbase
  7. Alexander Schnitzler commented on patch set 3 09/25/26, 18:43

    This change is ready for review.
  8. core-ci voted on patch set 3 V +1 09/25/26, 18:49 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/109294

  9. Stefan Bürk commented on patch set 3 CR -1 09/25/26, 19:08

    Reviewing ...
  10. Stefan Bürk commented on patch set 3 CR 0 09/25/26, 19:32

    Reply

    don't find the time yet.
  11. Oliver Klee voted on patch set 3 V +1 09/25/26, 20:08

  12. Benni Mack commented on patch set 3 09/30/26, 06:22

    BehaviourConformanceTest.php:75

    we now have an inconsistency in this test suite: TYPO3_REQUEST is FE, and ConfigurationManager is BE.
  13. Alexander Schnitzler uploaded patch set 4 V +1 10/02/26, 12:19

  14. core-ci voted on patch set 4 V +1 10/02/26, 12:27 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/109932

  15. Alexander Schnitzler rebased onto patch set 5 V +1 10/08/26, 09:54

  16. core-ci voted on patch set 5 V +1 10/08/26, 10:05 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/110476

  17. Oliver Klee voted on patch set 5 V +1 10/08/26, 12:38

  18. Alexander Schnitzler commented on patch set 5 10/09/26, 10:24

    BehaviourConformanceTest.php:75 Reply

    True, has been changed to FE context only.
  19. Alexander Schnitzler commented on patch set 5 10/09/26, 10:26

    btw: This is a preparation change for bigger refactorings in the Extbase persistence like the introduction of doctrine/orm (https://review.typo3.org/c/Packages/TYPO3.CMS/+/96533) or other performance improvements (currently not on gerrit). But without a better testsuite, we are acting rather blind. This change and the ones to come are boring but maybe a little context shows the goal a bit better.
  20. Benni Mack voted on patch set 5 CR +1 V +1 10/12/26, 19:03

  21. Georg Ringer voted on patch set 5 CR +1 V +1 10/12/26, 19:14

  22. Simon Schaufelberger voted on patch set 5 CR +1 V +1 10/13/26, 13:38

  23. Benni Mack voted on patch set 5 CR +2 V +2 10/14/26, 16:56

  24. Benni Mack merged patch set 6 10/14/26, 16:56