Merge conflict Gerrit cannot merge this change. It needs a rebase onto main.

Review main Ready for review Quiet

95867[FEATURE] Render pages without TypoScript

Oli Bartsch CR -2 V +1 Expert +2427 −49 updated 10/23/26, 13:36

Commit message

A site set now brings the Fluid templates it owns along, in a `views`
directory next to its `config.yaml`. The directory is found by its
name, the same way a set is not asked where its `labels.xlf` or its
`page.tsconfig` are, so nothing declares it. `PAGEVIEW` uses the page
templates in `pages` and the shared layouts and partials in `layouts`
and `partials`

Layouts and partials are shared between the scopes that render with
them, and a scope gets a sub directory of its own on top of them:
PAGEVIEW looks into `layouts/pages` and `partials/pages` before
`layouts` and `partials` of the same set. The planned `content` scope
will use `layouts/content` and `partials/content` in the same way,
keeping page and content element templates of the same name apart.

A set spells its directories in lower case and its files with a lower
case first letter, while a template root of "paths." keeps the
established upper case. Both spellings can appear in one installation
without either side having to know about the other: when a set brings
templates along, PAGEVIEW asks Fluid for the template of the page
layout starting lower case, and Fluid falls back to the upper case
one. Without such a set, the template is looked up as before.

Templates somewhere else stay a matter of the "paths." option of
PAGEVIEW, which keeps working exactly as it did and keeps the last
word. A set whose templates cannot move therefore configures them the
way it does today, and nothing about it has to change.

SetDefinition carries the directory a set is rooted in as "basePath",
derived from where its config.yaml lies instead of from where its
"typoscript" points, so that a set relocating its TypoScript does not
relocate its templates along with it. "Configuration" becomes a
declared package resource path, since templates may now legitimately
live there.

A set does not need an extension around it either. A project brings
sets along itself, in a "sets" directory next to the "sites" directory
of its configuration, found the way a site configuration is found.
Such a set is rooted where its config.yaml lies, which is an absolute
path below the project path and therefore resolvable everywhere an
"EXT:" path is. These sets are collected after the sets of all
packages, so a set of the project wins over a set of the same name
that an extension ships, and a project can take over a set it does not
own without patching the extension that provides it.

For page type 0, the frontend renders with an implicit PAGE object as
long as no PAGE object of TypoScript claims that type and a set of the
site brings templates along. A PAGE object of another page type leaves
page type 0 to it, whatever that object happens to be called: a
"page = PAGE" carrying a typeNum of its own renders that type and
hands neither the decision nor its properties to the implicit object.
The properties of a bare "page" are kept, so the implicit object is
configured without being declared, and "TypoScript > Active TypoScript"
points out that it renders the pages.

Where no page can be rendered at all, the frontend names the reason
instead of reporting a PAGE object nobody meant to write: a "page"
object occupied by another content object, or sets that bring no
"views/pages" directory along.

The paths of all sets of a site are collected in set dependency order,
so a set overrides the templates of the sets it depends on without
repeating their paths, file by file. Sets without a dependency
relation between them are ordered the way the site lists them in its
"dependencies", so that the site configuration and not the set name
decides which of them wins.

Test this change

Run a command in your clone of the TYPO3 Core. Then try what the change does and vote on Verified in Gerrit. How to review a patch

Check out patch set 15
git fetch https://review.typo3.org/Packages/TYPO3.CMS refs/changes/67/95867/15 && git checkout FETCH_HEAD
Cherry-pick onto main
git fetch https://review.typo3.org/Packages/TYPO3.CMS refs/changes/67/95867/15 && git cherry-pick FETCH_HEAD

Review

  • Code review CR -2 Blocked
    • Benni Mack -2
    • Matthias Vogel +1
  • Verified V +1 Needs a vote
    • Matthias Vogel +1
    • core-ci +1 CI

Relation chain 3

  1. main 97001[TASK] Give custom package resource definitions precedence over defaults Based on this change Matthias Vogel CR 0 V 0 Easy +43 −1
  2. 95867 [FEATURE] Render pages without TypoScript This change

  3. main 96900[!!!][TASK] Apply site sets in the order a site declares them This change is based on it Merged Oli Bartsch CR +2 V +2 Normal +173 −29

Issues 1

Files 80

Show 60 more files

Review log 2136

Show 711 earlier messages
  1. Oli Bartsch uploaded patch set 1 09/15/26, 21:49

  2. core-ci voted on patch set 1 V -1 09/15/26, 21:57 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/108642

  3. Oli Bartsch uploaded patch set 2 V -1 09/15/26, 22:55

  4. core-ci voted on patch set 2 V +1 09/15/26, 23:03 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/108643

  5. Georg Ringer voted on patch set 2 CR +1 V +1 09/16/26, 05:12

  6. Benjamin Franzke commented on patch set 2 09/16/26, 06:10

    Feature-110716-SiteSetTemplateRootPaths.rst:22

    I think the naming is not "clear" enough, in the sense whether this is about page or content templates. I'd actually love to see zero-configuration as for labels, typoscript and so on. That means: * Autoloading of `PAGEVIEW` paths **from the set folder** like `Sets/{myset}/pageview/layouts`, `Sets/{myset}/pageview/pages` and `Sets/{myset}/pageview/partials/` * Optional(!) possibility to define a template path outside the set (like proposed in this case), but with explicit naming, e.g. `pageViewPath` Explanation: Sets – similar to content blocks – were designed to work without extension/self references, allowing them to be drop-ins without the requirement to adapt self-references and extension keys and allowing them to live independent from extensions at some point (e.g. in the project instead of an extension).

    Feature-110716-SiteSetTemplateRootPaths.rst:30

    I think we need to declare the `PAGEVIEW` object implicitly if at least one set uses the `pageViewPaths`/`templates` feature. Having to declare this dependency feels redundant and incorrect, given that the template path support is added as first class set configuration qualifier.
  7. Oli Bartsch uploaded patch set 3 CR +1 V +1 09/16/26, 10:07

  8. Oli Bartsch commented on patch set 3 09/16/26, 10:08

    Feature-110716-SiteSetTemplateRootPaths.rst:22 Reply

    Agreed on all three points. Naming: pageViewPaths, so it says which templates it is about. Plural because template roots are an ordered list, like PAGEVIEW's own "paths.", so it takes a single path or a list. Zero configuration: a "pageview" directory next to config.yaml, found by name like labels.xlf and page.tsconfig, with lowercase pages/, layouts/ and partials/ as you sketched. Declared paths are optional and layer on top of it. Your drop-in point convinced me on a detail I had decided the other way before: a relative pageViewPaths value now resolves against the set, not the public directory, so a set never has to name its own extension key. One caveat towards real drop-ins: asset publishing runs per package, and which directories are public is declared in the extension's Configuration/Resources.php. A set's assets can be published from there, but the declaration lives outside the set and has to name its path, so a set still cannot bring CSS and JS along on its own. Probably worth a separate look, not something for this change.

    Feature-110716-SiteSetTemplateRootPaths.rst:30 Reply

    I actually had this in an earlier revision and dropped it again, because the obvious condition (no PAGE object anywhere in setup) breaks on core itself: EXT:seo declares "seo_sitemap = PAGE" with typeNum 1533906435, so adding the typo3/seo-sitemap set would kill type 0 rendering with "No page configured for type=0". Your formulation fixes that. The implicit object now applies when no PAGE object claims the *requested* type, which findPageObject() already determines, so a PAGE object of another type is irrelevant. It also bails out when "page" carries a value of its own ("page = TEXT"), and Active TypoScript shows an infobox that pages are rendered without a PAGE object, since the object itself is not listed anywhere. The typo3/page-view set is removed, it is redundant now.
  9. core-ci voted on patch set 3 V -1 09/16/26, 10:22 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/108668

  10. Oli Bartsch uploaded patch set 4 V -1 09/16/26, 11:05

  11. core-ci voted on patch set 4 V +1 09/16/26, 11:12 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/108674

  1. Oli Bartsch commented on patch set 4 09/16/26, 13:26

    Feature-110716-SiteSetTemplateRootPaths.rst:22 Reply Unresolved

    .

    Feature-110716-SiteSetTemplateRootPaths.rst:30 Reply Unresolved

    .
  2. Oli Bartsch uploaded patch set 5 V +1 09/16/26, 14:23

  3. core-ci voted on patch set 5 V +1 09/16/26, 14:31 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/108696

  4. Oli Bartsch updated the commit message in patch set 6 V +1 09/18/26, 09:06

  5. core-ci voted on patch set 6 V +1 09/18/26, 09:15 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/108798

  6. Oli Bartsch uploaded patch set 7 V +1 09/18/26, 10:41

  7. core-ci voted on patch set 7 V +1 09/18/26, 10:49 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/108801

  8. Oli Bartsch uploaded patch set 8 V +1 10/19/26, 11:13

  9. core-ci voted on patch set 8 V +1 10/19/26, 11:28 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/111622

  10. Oli Bartsch uploaded patch set 9 V +1 10/19/26, 11:36

  11. core-ci voted on patch set 9 V +1 10/19/26, 11:45 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/111626

  12. Oli Bartsch uploaded patch set 10 V +1 10/20/26, 07:30

  13. core-ci voted on patch set 10 V +1 10/20/26, 07:39 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/111751

  14. Oli Bartsch uploaded patch set 11 V +1 10/20/26, 09:01

  15. core-ci voted on patch set 11 V -1 10/20/26, 09:11 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/111771

  16. Oli Bartsch uploaded patch set 12 V -1 10/20/26, 09:31

  17. core-ci voted on patch set 12 V +1 10/20/26, 09:40 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/111779

  18. Matthias Vogel uploaded patch set 13 V +1 10/22/26, 11:48

  19. Matthias Vogel uploaded patch set 14 10/22/26, 11:58

  20. core-ci voted on patch set 13 V -1 10/22/26, 12:04 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/112097

  21. core-ci voted on patch set 14 V -1 10/22/26, 12:09 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/112100

  22. Matthias Vogel uploaded patch set 15 V -1 10/22/26, 12:38

  23. core-ci voted on patch set 15 V +1 10/22/26, 12:55 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/112104

  24. Matthias Vogel voted on patch set 15 CR +1 V +1 10/22/26, 12:58

  25. Benni Mack commented on patch set 15 CR -2 10/23/26, 13:36

    blocking until further notice (no offense)