Review main Merged

94177[TASK] Revert "[TASK] Include current site in cache tag collection"

Benjamin Franzke CR +2 V +2 Easy +0 −2 updated 05/30/26, 07:51

Commit message

This reverts commit b390dbe83e08ec37e731950c94ef6740ac35d1e0.

The reverted change introduced a regression by using a plain site
identifier as a cache tag. Site identifiers may contain characters
that are prohibited for cache tags, if the site was manually
created.

Technically these invalid characters are not allowed to
be created via the TYPO3 backend, but TYPO3 currently allows
to use these sites, if they were manually created.

The change was a factual FEATURE and therefore it is reverted
since this cache tag should only be added in TYPO3 v15 when
TYPO3 core ensures that only valid site identifiers are allowed
to be used.

Existing installations can easily add such a tag within a middleware
that adds a cache tag to the cache data collector, or via TypoScript:

```
page = PAGE
page {
  1779954161 = TEXT
  1779954161.stdWrap.addPageCacheTags.wrap = site_|
  1779954161.stdWrap.addPageCacheTags.data = site:identifier
}
```

This example has the same problem as the reverted patch, that
invalid site identifiers will cache invalid cache tags, but this
example is here to prove that the patch itself was not needed
to be backported and can therefore be reverted without replacement.

Review

  • Code review CR +2 Done
    • Helmut Hummel +2
    • Garvin Hicking +1
    • Anja Leichsenring +1
    • Christian Kuhn +1
    • Josua Vogel +1
    • Oliver Hader +1
  • Verified V +2 Done
    • Helmut Hummel +2
    • Garvin Hicking +1
    • Anja Leichsenring +1
    • Josua Vogel +1
    • Oliver Hader +1
    • core-ci +1 CI

Issues 1

Files 1

Review log 2225

  1. Benjamin Franzke uploaded patch set 1 05/28/26, 07:56

  2. core-ci voted on patch set 1 V +1 05/28/26, 08:04 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/102513

  3. Josua Vogel voted on patch set 1 CR +1 V +1 05/28/26, 08:05

  4. Oliver Klee updated the commit message in patch set 2 CR +1 V +1 05/28/26, 08:11

  5. Josua Vogel voted on patch set 2 CR +1 V +1 05/28/26, 08:13

  6. core-ci voted on patch set 2 V +1 05/28/26, 08:21 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/102514

  7. Garvin Hicking commented on patch set 2 CR +1 V +1 05/28/26, 08:38

    Thanks!
  8. Helmut Hummel voted on patch set 2 CR +1 V +1 05/28/26, 08:53

  9. Soren Malling voted on patch set 2 CR +1 V +1 05/28/26, 09:31

  10. Stefan Bürk voted on patch set 2 CR +1 V +1 05/28/26, 10:33

  11. Helmut Hummel commented on patch set 2 05/28/26, 10:52

    Here is what I would write instead: ``` page { 1779954161 = TEXT 1779954161.addPageCacheTags.wrap = site_| 1779954161.addPageCacheTags.data = site:identifier } ```
  12. Benjamin Franzke commented on patch set 2 05/28/26, 11:05

    Reply

    Thx for you recommendation. Please see: https://docs.typo3.org/permalink/t3tsref:confval-text-stdwrap > stdWrap properties are available on the very root level of the object. *This is non-standard!* You should use these stdWrap properties consistently to those of the other cObjects by accessing them through the property "stdWrap".
  13. Christian Kuhn voted on patch set 2 CR +1 05/28/26, 13:03

  14. Helmut Hummel commented on patch set 2 05/29/26, 07:42

    Reply

    Okay. I understand this recommendation, although I don't necessarily share the concerns about root level usage. I would still use this then: ``` page { 1779954161 = TEXT 1779954161.stdWrap.addPageCacheTags.wrap = site_| 1779954161.stdWrap.addPageCacheTags.data = site:identifier } ``` I'd prefer TEXT over COA, and data and wrap over insertData, because insertData is a tricky thing: https://gist.github.com/pgampe/3023546 and the COA container functionality is not required here.
  15. Oliver Hader voted on patch set 2 CR +1 V +1 05/29/26, 08:15

  16. Oliver Hader updated the commit message in patch set 3 CR +1 V +1 05/29/26, 08:59

  17. Oliver Hader commented on patch set 3 CR +1 V +1 05/29/26, 09:00

    Reply

    Thanks for the input. I took the freedom to use Helmut's suggestion to get move on with the revert of the underlying regression.
  18. core-ci voted on patch set 3 V +1 05/29/26, 09:07 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/102597

  19. Josua Vogel voted on patch set 3 CR +1 V +1 05/29/26, 10:43

  20. Anja Leichsenring voted on patch set 3 CR +1 V +1 05/29/26, 10:59

  21. Garvin Hicking voted on patch set 3 CR +1 V +1 05/29/26, 11:19

  22. Christian Kuhn voted on patch set 3 CR +1 05/29/26, 12:21

  23. Helmut Hummel voted on patch set 3 CR +1 V +1 05/30/26, 07:50

  24. Helmut Hummel voted on patch set 3 CR +2 V +2 05/30/26, 07:51

  25. Helmut Hummel merged patch set 4 05/30/26, 07:51