89845[BUGFIX] Add missing page type to settings
Commit message
Review
-
Code review CR +2 Done
- Benjamin Franzke +2
- Garvin Hicking +1
-
Verified V +2 Done
- Benjamin Franzke +2
- Garvin Hicking +1
- core-ci +1 CI
Issues 1
Files 11
- Build/Sources/Sass/component/forms/_input-group.scss+6 −1
- Build/Sources/Sass/component/_settings.scss+1 −1
- Build/Sources/TypeScript/backend/settings/type/page.tsnew+100 −0
- Build/Sources/TypeScript/backend/element-browser.ts+6 −2
- Build/types/TYPO3/index.d.ts+3 −0
- typo3/sysext/backend/Classes/Controller/BackendController.php+1 −0
- typo3/sysext/backend/Resources/Private/Language/locallang_settingseditor.xlf2+4 −0
- typo3/sysext/backend/Resources/Public/Css/backend.css+8 −7
- typo3/sysext/backend/Resources/Public/JavaScript/settings/type/page.jsnew+13 −0
- typo3/sysext/backend/Resources/Public/JavaScript/element-browser.js+1 −1
- typo3/sysext/core/Classes/Settings/Type/PageType.php3new+60 −0
Review log 2125
-
Benjamin Kott uploaded patch set 1 06/24/25, 14:19
-
core-ci voted on patch set 1 V -1 06/24/25, 14:34 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/80739
-
Benjamin Kott uploaded patch set 2
V -107/02/25, 05:23 -
core-ci voted on patch set 2 V +1 07/02/25, 05:30 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/81228
-
Benjamin Franzke uploaded patch set 3
V +107/02/25, 06:29 -
core-ci voted on patch set 3 V +1 07/02/25, 06:36 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/81232
-
Simon Schaufelberger commented on patch set 3 07/02/25, 12:14
-
Christian Kuhn commented on patch set 3 07/02/25, 12:45
that's imho a feature that is hidden away as task: this adds a new type. they are (sparsely) documented in the initial feature.rst and the docs team did a great job in https://docs.typo3.org/m/typo3/reference-coreapi/main/en-us/ApiOverview/SiteHandling/SiteSettingDefinitions.html documenting all types. this patch needs documentation which imho turns it into a feature. consider an extension uses the new type: it would trigger a dependency to the latest v13 patch level release since the solution would most likely fail otherwise. this, together with the situation that a .rst file is needed imho clearly make this change a new feature. -
Stefan Bürk commented on patch set 3 07/02/25, 12:46
Reply
Had the same thoughts testing the change. -
Benjamin Franzke commented on patch set 3 07/02/25, 13:14
Reply
> consider an extension uses the new type: it would trigger a dependency to the latest v13 patch level release since the solution would most likely fail otherwise Soft failure, yes (The set that is uses an undefined type is invalidated, creates log entries and site settings and site overview will show respective errors to integrators) Which is actually fine… an extension that uses the `page` settings type "only" needs to require the respective v13 version as minimum requirement. Of course this is debatable, but this is not a feature in our opinion: We classify it as a completion of the existing site settings feature (which is actually a *feature*, not just extended functionality) of TYPO3 v13, based on feedback that demanded to use this type in extensions since type `int` does not provide good user guidance. I agree that we need to update documentation, which we'll take care of. We didn't enumerate all types in the initial site sets Feature.rst, therefore there is nothing to update. It's also not worth an Important.rst (nothing every integrator needs to read), but we'll be in contact with documentation team and update this. Please either block this change (-2) if you'd like to set a hard blocker or resolve the comment, so we know how to continue. Thanks for you input!locallang_settingseditor.xlf:39Replygrouping of type-related values. -
Christian Kuhn commented on patch set 3 07/02/25, 14:28
Reply
> Soft failure, yes (The set that is uses an undefined type is invalidated, creates log entries and site settings and site overview will show respective errors to integrators) that's good. > Which is actually fine… an extension that uses the `page` settings type "only" needs to require the respective v13 version as minimum requirement. which is another sign for a feature. > Of course this is debatable, but this is not a feature in our opinion: > We classify it as a completion of the existing site settings feature (which is actually a *feature*, not just extended functionality) of TYPO3 v13, based on feedback that demanded to use this type in extensions since type `int` does not provide good user guidance. it's 3/4 year since 13.4.0. how long should we keep adding features? > I agree that we need to update documentation, which we'll take care of. We didn't enumerate all types in the initial site sets Feature.rst, therefore there is nothing to update. Invalid point: The fact that the different types were not properly enumerated in the initial feature.rst was sloppy core documentation in the first place. Using this as "and because of this incompleteness, a new feature is now not a feature" does not cut it. The types are clearly API, they are supposed to be used by devs and integrators, they need proper documentation in the changelogs. Remember the changelog rst files are not only targeted towards the documentation team as a base for docs, but have various target groups (marketing, project devs looking for details, reading up details in various years when upgrading, ...). We're adding a new spectacular solution here that has communication needs. We loose this when we mask a feature and task and skip required docs: devs don't know, marketing does not know, documentation team does not know (they had to make something up for the initial types already, according to github, 4 docs team members were involved, no core team member), upgraders don't know about why and when. This new type clearly is worth being properly documented in the core changelog files, and we should actually extend the main feature.rst with sections about the single types. > Please either block this change (-2) if you'd like to set a hard blocker or resolve the comment, so we know how to continue. Thanks for you input! No. It's your responsibility to take the decision, either by bending rules in this case, or not. I will neither block nor give you the blessing to do so. I'm unsure if this is *really* so very much required in v13 to force bending rules, you know this area better than me and have deeper insight into early adopter uses, impact on core version restrictions and possible side effects of this within the site settings editor. I'm only stating you are bending rules if this feature is merged as task and outline possible negative side effects. I think backporting a feature into a stable release should be a conscious decision that is clearly stated. That's why I raised the comment. -
Benjamin Kott commented on patch set 3 07/03/25, 06:30
Reply
I think it's fine, and we should talk about such things. This was indeed discussed, and the need was expressed to close the gaps in the Settings API that did not make it in time. It feels necessary and our responsibility not to let adopters sit in an unfinished state. While we promote the usage of sets and are "happy" with the general acceptance, we also encounter issues that make it hard to use. If we compare this example of the page type—it was requested from the beginning but did not make it in time. To explain the problem we are trying to address with this patch: The constant editor was always used next to a page tree. This made it at least possible to find the page or storagePIDs while editing the constants. This changed with the Settings Editor, as it's running in a different scope. Now you really need to have a second window open or write the IDs down—because you cannot quickly peek at the page tree. While we were aware that this was missing, it only really became a visible problem when we started intensively using the settings ourselves. There are more issues we've noted that feel essential to add—so users don't feel downgraded when switching to the new API. This includes validation rules or additional configuration per setting for simple things like "min" or "max" values. While this can technically be described as a feature, it feels like the bare minimum required to use the initial feature in the first place. At least to have feature parity with the Constant Editor, which we are sad to say is not currently the case. Even providing "real" features like implementing the Settings API is still challenging, and we discussed a lot how we can mitigate the issue and help developers in using and adopting the system. While we agree that we will not backport the usages we are currently working on bringing to core for v14, like the dashboard settings... We also need to keep in mind how people are moving from v13 to v14 and support extensions that target both versions. Since we are finding a lot through actual usage, keeping the stable version stable remains the top priority. But to achieve that, even with the more complex patches, the best way seems to be keeping things in sync as long as possible and bringing all necessary changes to v13 as well. It was discussed, for example, whether the page type patch is a feature or not—do we need to inform the users via an important RST or not? For now, we decided against that and opted instead to adjust the documentation, as the initial types were also not documented explicitly. So the decision is: do we try to keep the Settings stable across multiple versions, or do we need to tell users to stop using settings and sets and come back for them in v14? -
Benjamin Kott updated the commit message in patch set 4
CR +1V +107/04/25, 07:58 -
core-ci voted on patch set 4 V +1 07/04/25, 08:06 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/81313
-
Benjamin Franzke voted on patch set 4 CR +1 V +1 07/04/25, 08:08
-
Benjamin Franzke commented on patch set 4 07/07/25, 13:50
Reply
> No. It's your responsibility to take the decision, either by bending rules in this case, or not. Alright, Benji updated the commit message and explained our reasoning for enhancing this v13 feature to bring it on par with the previous user experience. Documentation will be adapted once merged/adapted. Thanks. -
Garvin Hicking commented on patch set 4 07/07/25, 19:36
If we want to document this new type on the TYPO3 documentation pages, we could need an "Important.rest" file for this; else it's possible the docs team will not notice this, because no "Changelog-to-doc" workflow will be triggered. If you agree this is ok I'm offering to write the ReST file, of course. -
Simon Schaufelberger wrote 07/07/25, 19:39
Removed cc Simon Schaufelberger. -
Garvin Hicking commented on patch set 4 CR +1 V +1 07/07/25, 19:42
PageType.php:27Are there plans to enable this type in the core's site set definitions at some later point? Like felogin.pid? -
Benjamin Franzke commented on patch set 4 07/08/25, 07:02
Reply
Just for my understanding: What's wrong in providing a PR to the documentation manually? Isn't it documentable as well then? Do we really *need* a RST to document things, I thought it would also be fine if we just file a PR to the documentation manually… I think, if we start adding types to the changelog (which should have been listed there from the start – that's my bad), we should basically list all types, including the new page type in the initial Feature.rst. I just don't want to spam integrators with Important.rst's that are not *that* important for everyone to read. This is basically something people will need once they adapt to use settings instead of TypoScript constants. -
Benjamin Franzke commented on patch set 4 07/08/25, 07:09
PageType.php:27Replyyes, we could have actually done that right away… (existing string value would automatically be converted to int and therefore still work) -
Garvin Hicking commented on patch set 4 07/08/25, 07:47
Reply
Sure, providing a PR for the docs alongside works for me/us as well. I didn't see that this is the plan in the commit message, so I just wanted to raise this issue. -
Benjamin Franzke commented on patch set 4 CR +2 V +2 07/08/25, 07:50
PageType.php:27ReplyWe'll do that later. -
Benjamin Franzke merged patch set 5 07/08/25, 08:01
-
Benjamin Franzke commented on patch set 5 07/08/25, 08:57
locallang_settingseditor.xlf:39