WIP|POC:Out of the inline relation rendering for backend #73

DanielSiepmann · 2024-01-09T16:51:23Z

This is a PoC, not final implementation.

Relates: #68

DanielSiepmann · 2024-01-09T16:51:40Z

I opened this PR so we have an easy way to discuss stuff based on the PoC :)

Classes/Backend/Preview/PreviewRenderer.php

DanielSiepmann · 2024-01-09T16:55:27Z

Classes/DataProcessing/ContentBlockDataResolver.php

+                $processedField = $this->processField(
+                    $tcaFieldDefinition->getFieldType(),
+                    $processedField,
+                    $tcaFieldDefinition,
+                    $table,
+                    $depth,
+                    $context
+                );
+                if (isset($processedField['__resolvedGrid'])) {
+                    $processedContentBlockData[$tcaFieldDefinition->getIdentifier() . '_grid'] = $processedField['__resolvedGrid'];
+                    $processedField = $processedField['__resolvedField'];
+                }


This whole process feels a bit clunky to me. But I couldn't figure out a better way yet …
Maybe you have ideas, or we can put our heads together.

I moved the Result into a dedicated struct and added _grids to ContentBlockData. This way it is better separated from "normal" processed fields.

DanielSiepmann · 2024-01-09T16:55:59Z

Classes/DataProcessing/GridFactory.php

+                $context,
+                $column,
+                $record->_raw,
+                // $tableName,


Can be done once we have a new TYPO3 release with https://review.typo3.org/c/Packages/TYPO3.CMS/+/82411

Yeah, perfect!

DanielSiepmann · 2024-01-09T17:00:50Z

Classes/Backend/Preview/PreviewRenderer.php

        $data = $contentBlockDataResolver->buildContentBlockDataObjectRecursive(
            $contentElementDefinition,
            $contentElementTableDefinition,
            $record,
-            $contentElementTable
+            $contentElementTable,
+            0,
+            $item->getContext(),
        );


There are way to many arguments for my taste.
Maybe we can come up with some other API, encapsulation or something else.

Mhh, yeah slowly. Could be moved into a single carrier class. Let's first get this done and refactor this at the very end.

nhovratov · 2024-01-09T21:19:21Z

So, I played around a bit and it already works really well. What disturbed me was the dragging possibility was still there for the children. This can't work of course. So I decided to copy the relevant partials and adjust them a bit for our purpose. I think this is ok. When there are changes in the future in this area, the Core is forced to adjust these partials as well, as this is part of the Core then.

You can use it like this now:

<f:render partial="PageLayout/Grid" arguments="{data: data, identifier: 'children', allowEditContent: 1}"/>

ContentBlockData must be provided, as it holds all the grids. Then the identifier of the Relation field as a string. Lastly allowEditContent. This should possibly be just enabled by default.

DanielSiepmann · 2024-01-10T07:18:54Z

Classes/DataProcessing/ContentBlockData.php

+        if ($name === '_grids') {
+            return $this->_grids;
+        }


I like that. The only drawback: We now have this always empty in frontend.
Maybe we can adjust that and use this to provide already rendered grids for frontend? I guess EXT:gridelements and EXT:container did the same?
That way it would be the "same" experience in frontend and backend.
Processed holds the data while _grids holds the ready to use stuff for rendering based on context.

We could alter the generation. Do what we do right now in case Context is provided. Use TypoScript rendering otherwise.
Or provide a callback from the outer scope to generate the grid. That way the resolver doesn't need to know inspect and act, but the caller can influence the result.

What do you think @nhovratov?

This sounds really interesting. This would allow users to alternate between standard rendering (TypoScript) and custom rendering via processed. Let's keep this idea in mind. We should think about a better naming then, I believe. Like "children" or "nestedContent". Idk. Of course this would only make sense for tt_content records, as other record types rarely have a TypoScript rendering definition, altough it would be possible.

DanielSiepmann · 2024-01-10T07:29:01Z

The current state has a small issue with the language flag icon within the grid. It misses some spacing:

But the only difference in templates is the drag and drop class.
But Firefox shows there is a missing whitespace which would create the space.

nhovratov · 2024-01-10T08:38:38Z

The current state has a small issue with the language flag icon within the grid. It misses some spacing:
But the only difference in templates is the drag and drop class. But Firefox shows there is a missing whitespace which would create the space.

Yeah, it's the whitespace. I've seen the same issue with EXT:container already. It should be fixed in the Core styling imo, so the space does not rely on space characters.

nhovratov · 2024-01-10T18:20:05Z

FYI Patch for the gap: https://review.typo3.org/c/Packages/TYPO3.CMS/+/82427

nhovratov · 2024-01-23T21:17:35Z

Current state with patches applied to Core

Nested Content Element (works without patches, as it is tt_content)
Generic Record Type (without a type field + StandardContentPreviewRenderer)
Generic Record Type (with type field + StandardContentPreviewRenderer)

DanielSiepmann · 2024-01-24T06:37:59Z

You are awesome @nhovratov, thanks for all the efforts :)

This is a PoC, not final implementation. Relates: #68

nhovratov · 2024-02-17T17:28:39Z

The backend preview feature for nested content elements is included in this commit: 2779792
Due to refactoring in the meantime the merge conflicts got too complex, so I had to implement this again outside of this PR.

Two things left to do:

Support for preview of Collections other than tt_content. See Core patch: https://review.typo3.org/c/Packages/TYPO3.CMS/+/82442/13
Pre-rendered nested Content Elements in the frontend (like EXT:container does it).

I would say, let's open a new issue for that. This PR got a little messy over the time.

DanielSiepmann force-pushed the 68-grid-poc branch from 9cd2f1f to c7fbf94 Compare January 9, 2024 16:54

DanielSiepmann commented Jan 9, 2024

View reviewed changes

DanielSiepmann force-pushed the 68-grid-poc branch from c7fbf94 to 9390ab1 Compare January 9, 2024 16:59

DanielSiepmann commented Jan 9, 2024

View reviewed changes

DanielSiepmann commented Jan 10, 2024

View reviewed changes

nhovratov force-pushed the 68-grid-poc branch from cf5dd3e to c4a7321 Compare January 27, 2024 01:29

DanielSiepmann changed the title ~~WIP|POC:Out of the box grid rendering for backend~~ WIP|POC:Out of the inline relation rendering for backend Jan 30, 2024

DanielSiepmann and others added 9 commits February 3, 2024 19:21

WIP|POC:Out of the box grid rendering for backend

15d0354

This is a PoC, not final implementation. Relates: #68

[TASK] Add strict comparison

e6e7423

[TASK] Hold grids in dedicated grids array in ContentBlockData

78e88ee

[TASK] Render label of field in grid column header

1abea92

[TASK] Use proper DI for ContentBlockDataResolver

43b8d18

[TASK] Add trailing slashes for better debug output

484de03

[TASK] Enable allowEditContent if not explicitly disabled

b8e597e

[BUGFIX] Fix missing check for plain arrays in GridFactory

481a8c5

[TASK] Add out-commented PreviewRenderer for generic tables

80ff5a9

nhovratov force-pushed the 68-grid-poc branch from c4a7321 to 80ff5a9 Compare February 3, 2024 18:23

nhovratov force-pushed the main branch 2 times, most recently from 8991e76 to f4eef35 Compare February 9, 2024 22:56

nhovratov closed this Feb 17, 2024

nhovratov mentioned this pull request Feb 17, 2024

Pre-rendered Nested Content Elements #126

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

WIP|POC:Out of the inline relation rendering for backend #73

WIP|POC:Out of the inline relation rendering for backend #73

DanielSiepmann commented Jan 9, 2024

DanielSiepmann commented Jan 9, 2024

DanielSiepmann Jan 9, 2024

nhovratov Jan 9, 2024

DanielSiepmann Jan 9, 2024

nhovratov Jan 9, 2024

DanielSiepmann Jan 9, 2024

nhovratov Jan 9, 2024

nhovratov commented Jan 9, 2024

DanielSiepmann Jan 10, 2024

nhovratov Jan 10, 2024

DanielSiepmann commented Jan 10, 2024

nhovratov commented Jan 10, 2024

nhovratov commented Jan 10, 2024

nhovratov commented Jan 23, 2024

DanielSiepmann commented Jan 24, 2024

nhovratov commented Feb 17, 2024

WIP|POC:Out of the inline relation rendering for backend #73

WIP|POC:Out of the inline relation rendering for backend #73

Conversation

DanielSiepmann commented Jan 9, 2024

DanielSiepmann commented Jan 9, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

nhovratov commented Jan 9, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

DanielSiepmann commented Jan 10, 2024

nhovratov commented Jan 10, 2024

nhovratov commented Jan 10, 2024

nhovratov commented Jan 23, 2024

DanielSiepmann commented Jan 24, 2024

nhovratov commented Feb 17, 2024