Fix controlled inner blocks parent block attributes updates #35827
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
talldan
added
[Type] Bug
|
Looking at the diff here, there are some changes that make me worry. Let me dig deeper into the issue. |
|
Size Change: +28 B (0%) Total Size: 1.08 MB
ℹ️ View Unchanged
|
|
It should be almost exactly the same, the only difference is that if only the block wrapper had updates, that gets added to clientIds list now. My thinking is that this needs to bubble up to the parent entity (e.g. if the only the template part block wrapper is edited, that should still save to the template). |
youknowriad
reviewed
Oct 21, 2021
youknowriad
reviewed
Oct 21, 2021
talldan
force-pushed
the
fix/controlled-inner-blocks-wrapper-attributes-saving
branch
from
fdc41eb to
1242644
Compare
|
So I think what I've unearthed is that there are two separate bugs.
|
|
The failing tests indicate there are more issues with this fix too. @youknowriad any other ideas how to solve the problems? |
youknowriad
reviewed
Oct 21, 2021
| @@ -246,15 +246,20 @@ function updateParentInnerBlocksInTree( state, tree, updatedClientIds ) { | |||
| for ( const clientId of updatedClientIds ) { | |||
| let current = clientId; | |||
| do { | |||
| clientIds.add( current ); | |||
|
|
|||
| if ( state.controlledInnerBlocks[ current ] ) { | |||
| controlledParents.add( current ); | |||
There was a problem hiding this comment.
I guess if we avoid this on the first level of the loop (when clientId is part of updatedClientIds) to avoid marking its innerBlocks as dirty. which should resolve your second point maybe ?
There was a problem hiding this comment.
Maybe just moving this line to the check line 256?
|
Both of your points 1 and 2 are logical to me. That said, I feel the implementation we have is probably too complex for nothing. I'll see if there's a way to make it a bit more simple. |
youknowriad
force-pushed
the
fix/controlled-inner-blocks-wrapper-attributes-saving
branch
from
f108d94 to
3e8fa4d
Compare
|
Ok, I think what's confusing is that sometimes the function is called with the actual blocks that were modified (attributes), but sometimes it's called with their parents, so it's hard to reason about, I'm thinking we should add another argument to separate these two use-cases. |
|
I think I remove the ambiguity with the last update, let's see. |
13 tasks
|
Thanks for iterating on this. Still some failing tests I see. I'll try to give this a bit more testing on Monday to see what the issue is there. |
38 tasks
talldan
commented
Oct 25, 2021
| @@ -1980,7 +1980,7 @@ export const __experimentalGetPatternTransformItems = createSelector( | |||
| if ( | |||
| blocks.some( | |||
| ( { clientId, innerBlocks } ) => | |||
| innerBlocks.length || | |||
| innerBlocks?.length || | |||
There was a problem hiding this comment.
@youknowriad This should hopefully resolve the failing tests. 🤞
It looks like with the recent change, some blocks in state now won't have the innerBlocks property. I think the reason is that if a block only ever gets updated using the false option it won't ever reach this line of code that adds the property (only the parents will):
https://github.com/WordPress/gutenberg/pull/35827/files#diff-3c4b969f7547d02bc04a2c554e0f7d21da07c3f779714acf02800ee0341b50d9R276-R280
The alternative fix would be to ensure the innerBlocks property is present, but that might involve changing the updateParentInnerBlocksInTree again. Let me know what you think.
edit: also updated the failing unit test in 14bcacf, that had the same issue.
There was a problem hiding this comment.
Seems like there's also this code that expects innerBlocks to be defined, which is causing the remaining failing test:
gutenberg/packages/blocks/src/api/serializer.js
Lines 265 to 283 in d209b27
Perhaps it's better to look at fixing the root cause.
There was a problem hiding this comment.
yes, It seems we should consider adding an empty inner blocks in some init code somewhere to avoid that. Some third-party code could also rely on it.
There was a problem hiding this comment.
@youknowriad Did some more staring at this problem, and I think the cause is the UPDATE_BLOCK case, which isn't passing in a valid tree to updateParentInnerBlocksInTree. Hopefully the change in 04d1853 is ok.
We could also add some initialization code, but I don't think it's necessary right now.
talldan
force-pushed
the
fix/controlled-inner-blocks-wrapper-attributes-saving
branch
from
14bcacf to
04d1853
Compare
youknowriad
approved these changes
Oct 29, 2021
talldan
deleted the
fix/controlled-inner-blocks-wrapper-attributes-saving
branch
|
Thanks for your help with this one, glad we could get it sorted! |
|
Thanks for catching and looking into this |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Fixes #35664
I bisected and found that the problem was introduced in #34241. I scanned the code and I think I spotted the issue:
It looks like the intent here was to traverse up and stop when reaching a block that has controlled inner blocks.
It has an unwanted side effect that if it's only the parent block (and not the controlled inner blocks) that changed the parent is also skipped from being added to the clientIds list.
Instead I think it needs to check the parent on each loop, and only traverse up if the parent isn't a controlled inne rblock
How has this been tested?
Types of changes
Bug fix (non-breaking change which fixes an issue)
Checklist:
*.native.jsfiles for terms that need renaming or removal).