Review main Merged

91709[BUGFIX] Use correct setting type for page selections in sets

Georg Ringer CR +2 V +2 Easy +4 −4 updated 08/28/26, 09:42

Commit message

The setting type should be `page` instead of `int` to provide a page
selector to the user.

Review

  • Code review CR +2 Done
    • Georg Ringer +2
    • Benjamin Franzke +1
    • Stefan Bürk +1
  • Verified V +2 Done
    • Georg Ringer +2
    • Benjamin Franzke +1
    • core-ci +1 CI

Issues 1

Files 2

Review log 1723

  1. Georg Ringer uploaded patch set 1 11/17/25, 09:29

  2. Georg Ringer commented on patch set 1 11/17/25, 09:29

    This change is ready for review.
  3. core-ci voted on patch set 1 V +1 11/17/25, 09:37 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/89923

  4. Josef Glatz voted on patch set 1 CR +1 V +1 11/17/25, 09:42

  5. Benni Mack voted on patch set 1 CR +1 V +1 11/17/25, 11:01

  6. Garvin Hicking commented on patch set 1 11/17/25, 11:30

    IMO this is more a task than a bugfix. Currently no core (also no tests?) seem to have "type: page" set for Site definitions. Should we broadly use it everywhere instead of one patch for every occurence? Also: Does "default: 0" work properly with type:page, did you check? Because "0" might not be a valid default and we would maybe need to omit a default then? That would be easier if we had tests in place... ;-)
  7. Georg Ringer uploaded patch set 2 CR +1 V +1 12/29/25, 17:57

  8. Georg Ringer updated the commit message in patch set 3 12/29/25, 17:58

  9. core-ci voted on patch set 2 V +1 12/29/25, 18:11 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/92786

  10. core-ci voted on patch set 3 V +1 12/29/25, 18:13 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/92787

  11. Benni Mack wrote 04/21/26, 15:02

    Removed reviewer Benni Mack.
  12. Benni Mack updated the commit message in patch set 4 V +1 04/22/26, 06:54

  13. core-ci voted on patch set 4 V -1 04/22/26, 11:34 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/100827

  14. Benni Mack rebased onto patch set 5 V -1 05/04/26, 19:14

  15. core-ci voted on patch set 5 V +1 05/04/26, 19:23 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/101494

  16. Benni Mack rebased onto patch set 6 V +1 08/14/26, 12:34

  17. core-ci voted on patch set 6 V +1 08/14/26, 12:41 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/106699

  18. Stefan Bürk commented on patch set 6 CR +1 08/15/26, 18:07

    Unresolved

    Technically this is correct and safe. I verified the parts that were questioned in patch set 1: * default: 0 is fine with type: page. PageType::validate() returns true for any int, so SettingDefinitionValidation::validate() does not throw. PageType also does not implement SettingsTypeOptionAwareInterface, so no option validation kicks in. * The stored values do not change: IntType::transformValue() and PageType::transformValue() both end in (int)$value, and none of the four settings declares min/max/step. So there is no migration concern for existing config/sites/*/settings.yaml. * Nothing besides the settings editor branches on the type string - the only consumers are SettingsTypeRegistry and the three controllers resolving getJavaScriptModule(). Outside the backend the widget degrades to a plain number field, as designed in #106953. * Scope is complete: after this patch all single-page-uid settings shipped by core use page, and no int-typed page setting is left. The remaining int settings are genuine numbers. (felogin.pid, indexedsearch.rootPidList and seo.sitemap.pages.excludePagesRecursive are page lists typed as string - there is no list type for that yet, so that would be a separate feature.) Three things I would like to see before this gets a +2: 1. Please answer Garvin's patch set 1 comment and resolve the thread, especially the [BUGFIX] vs [TASK] point. It is not cosmetic: it decides whether the change belongs on 14.3 at all. I lean [TASK], because nothing malfunctions today - the editor widget and the declared semantics change. 2. Forge #108160 is titled "Use type page for page selection in EXT:indexed_search" and is filed under the Indexed Search category, but since patch set 2 the change also touches three felogin settings. Please widen the issue or split off a second one, otherwise the release notes under-report the change. 3. There is still no test for the page type at all - typo3/sysext/core/Tests/Unit/Settings/Type/ has tests for bool, color, int, number, stringlist and string, but no PageTypeTest, and no set fixture uses type: page. With this patch core becomes the first consumer of that type. A PageTypeTest modelled on IntTypeTest would be cheap and would settle the "does default: 0 work" question permanently. A follow-up is acceptable for me, but it should exist. Two adjacent documentation glitches I noticed while reviewing, feel free to pick them up here or leave them: indexed_search/Documentation/Configuration/TypoScript/Index.rst declares :Type: boolean for targetPid, and in felogin/Documentation/Configuration/TypoScript.rst the :Site set setting: of redirectPageLogout is empty while redirectDisable points at felogin.redirectPageLogout. Also: was dropping 13.4 from Releases: in patch set 4 a deliberate release-team decision? The page type was backported to 13.4 with #106953, so it would be technically possible there. +1 on the code itself; the open points above are about scope, issue reference and test coverage rather than correctness.

    Reply

    I share that reading. Nothing malfunctions today - an integrator can enter a page uid into the int field and it is stored and used correctly; what changes is the editor widget and the declared semantics of the setting. That makes it a [TASK] for me, and it matters beyond the tag, because it decides whether this belongs on 14.3 at all. On the "no core usage / no tests" part: that is still true after six patch sets. typo3/sysext/core/Tests/Unit/Settings/Type/ has tests for bool, color, int, number, stringlist and string but no PageTypeTest, and no set fixture uses type: page - so with this change core becomes the first consumer of a type that has never been exercised. I verified the behaviour by hand (PageType::validate() accepts any int so default: 0 validates, and transformValue() casts to int exactly like IntType, so stored values are unaffected), but that verification should live in a PageTypeTest rather than in a review comment.
  19. Georg Ringer commented on patch set 6 08/27/26, 17:27

    no matter if it is a bugfixx or a task, it should belong into 14.3 as well
  20. Benjamin Franzke commented on patch set 6 CR +1 08/28/26, 08:59

    Reply Unresolved

    I agree that one could discuss BUGFIX vs TASK here. I'd like to see it as a usability fix and therefore BUGFIX is fine. Explanation: A clear page input field is missing the functionality to select a page => bug. Therefore +1 for BUGFIX from my site.
  21. Benjamin Franzke voted on patch set 6 V +1 08/28/26, 09:04

  22. Georg Ringer voted on patch set 6 CR +2 V +2 08/28/26, 09:42

  23. Georg Ringer merged patch set 7 08/28/26, 09:42