#81036TCA l18n_parent is processed for sys_language_uid 0
Description
When opening a TCA Form of e.g. a tt_content record of sys_language_id 0 the l18n_parent field is processed, though the field isn’t even displayed.
If you v a lot of tt_content records on the same pid this leads to major performance issues.
E.g. if arround 45k tt_content records on the same pid which are used by a container record via IRRE (kind of like the news does).
If there r 10 Records attached to the container records that leads to 10×45k interations in TYPO3\CMS\Backend\Form\FormDataProvider\AbstractItemProvider\addItemsFromForeignTable which takes 40sec. On my maschine :).
The only way to get arround this is to set type=>none to l18n_parent via override.
This behavior did not exist in TYPO3 6.1 (not sure about later versions).
I assume it has been introduced by the rewriting of the whole tce/form stuff.
I think this is more a general issue, imho fields that r not displayed at all by e.g. DisplayCond should be filtered earlier so that dont make their way through the itemProvider.
Note: this issues is not restricted to tt_content. In general: fields are processed by itemProviders/Dataproviders no matter if they r acctually displayed/used or not.
Reviews 3
Related issues 5
- Bug #82730Opening record in backend painfully slow with foreign table and large main table
- Bug #84247TCA label_userfunc called for every record in current page/folder
- Bug #92863TCA change for sys_file_reference l10n_parent kills performance
- Bug #105186Slow record list and edit form in backend for tables with many records
- Bug #107147TCA inline calls label_userFunc too often
Files 3
History 1013
-
Andre Koller changed Related to #82730 <span data-processor="relativetime" data-value="2017-11-14T15:56:32+00:00" > 11/14/17, 15:56 </span>
-
Susanne Moog changed category <span data-processor="relativetime" data-value="2018-01-24T20:11:10+00:00" > 01/24/18, 20:11 </span>
-
Webadmin no-lastname-given changed Related to #84247 <span data-processor="relativetime" data-value="2018-03-13T16:54:37+00:00" > 03/13/18, 16:54 </span>
-
Markus Klein changed Related to #92863 <span data-processor="relativetime" data-value="2020-11-17T18:46:50+00:00" > 11/17/20, 18:46 </span>
-
Gerrit Code Review pushed patch set 1 for main to change 93973 status
NewUnder Review <span data-processor="relativetime" data-value="2026-05-05T07:22:10+00:00" > 05/05/26, 07:22 </span> -
Gerrit Code Review pushed patch set 1 for 14.3 to change 95549 <span data-processor="relativetime" data-value="2026-09-02T07:49:03+00:00" > 09/02/26, 07:49 </span>
-
Benni Mack commented and changed status
Under ReviewResolved done0 %100 % <span data-processor="relativetime" data-value="2026-09-02T08:00:13+00:00" > 09/02/26, 08:00 </span>Applied in changeset commit:57a9572cc03852c8ffba5057a5fd8ed45612c34b.
-
JAKOTA Design Group GmbH commented and changed attached Screenshot 2026-09-13 at 21.01.42.png attached Screenshot 2026-09-13 at 21.03.03.png <span data-processor="relativetime" data-value="2026-09-13T20:13:40+00:00" > 09/13/26, 20:13 </span>
Hi,
I’m experiencing this issues also on 13.4 and after applying this patch the backend works again as expected.
This is a IRRE ajax request before this patch:
Screenshot%202026-09-13%20at%2021.01.42.png
it takes 30s.
After applying this patch the same request only takes 411ms.
Screenshot%202026-09-13%20at%2021.03.03.png
Is there a reason why this patch was not considered for 13.4?
Thanks
-
Gerrit Code Review pushed patch set 1 for 13.4 to change 95809 status
ResolvedUnder Review <span data-processor="relativetime" data-value="2026-09-14T10:34:40+00:00" > 09/14/26, 10:34 </span> -
Georg Ringer changed Related to #107147 <span data-processor="relativetime" data-value="2026-09-28T20:16:47+00:00" > 09/28/26, 20:16 </span>
-
Georg Ringer changed Related to #105186 <span data-processor="relativetime" data-value="2026-09-28T20:20:31+00:00" > 09/28/26, 20:20 </span>
-
Benni Mack commented and changed status
Under ReviewClosed <span data-processor="relativetime" data-value="2026-10-23T08:52:32+00:00" > 10/23/26, 08:52 </span>We can close this issue I think.
What has happened in the past 10 years — 57a9572cc03, “[BUGFIX] Skip foreign table query for transOrigPointerField on default language” (Resolves #81036), merged in September.
The fix is in TcaSelectItems.php:68-70. For the translation-parent field on records in the default language or “all languages”, it skips the foreign-table query entirely.
The performance issue is a known issue. Eg. we had to change the TCA of news to change the parent selector to a group field, since we had 25k news records in default language, which is performance hell.
The part about processing fields depending on the display condition is not that easy. There was a lot of thinking already by Christian Kuhn about that.
Maybe get in touch with him on Slack for more internal details on the processing.