Fix editor crash when converting block with visible styles to reusable (after a save and page reload) #29059
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
[Feature] Synced Patterns
talldan
requested review from
ajitbohra,
ellatrix,
nerrad and
ntwb
as code owners
|
Size Change: +35 B (0%) Total Size: 1.4 MB
ℹ️ View Unchanged
|
talldan
force-pushed
the
fix/reusable-block-conversion-crash
branch
from
ad70bd4 to
a1cc4ac
Compare
|
Ok - I discovered that while my change fixed the problem I was trying to address, I'd made a mistake and introduced other bugs (thanks tests). I'll need to revisit this and find another way. |
|
This should be ready for review now, fixed by adding some extra guarding around the rendering of block styles. |
talldan
force-pushed
the
fix/reusable-block-conversion-crash
branch
3 times, most recently
from
38fa982 to
33de4b0
Compare
talldan
force-pushed
the
fix/reusable-block-conversion-crash
branch
from
33de4b0 to
9390e3c
Compare
|
yeah, that's the redux zombie bug, I thought we had it covered in useSelect (with the try/catch but maybe not in some cases) |
|
These things are very hard to debug, ideally the parent component should rerender first hide the whole inspector before the sub components and the "try catch" in useSelect is meant to actually catch these potential rerender issues when the child rerenders first. That said, I'm not sure I can provide more value, than an explanation of what's happening so I'm fine with the fix as is. It's a stop gap fix though and it can happen again if a new component is added there for instance. So the test is a very good addition. Thanks. |
youknowriad
approved these changes
Mar 8, 2021
|
Thanks for reviewing. I'll try to find out why the new test is failing. Hopefully it will just be a case of making it more robust. edit: the test needed to be updated for the new dialog that shows when creating a reusable block. |
|
This issue #29498 is also a zombie bug problem, I wonder if the current PR has an impact there. |
… prevent attempt to render non-existant block Refine dependencies Refactor a little bit more
talldan
force-pushed
the
fix/reusable-block-conversion-crash
branch
from
9390e3c to
2ffc493
Compare
|
It doesn't seem to affect #29498 (nav links don't have styles), but I also couldn't reproduce that one against the latest code in |
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 #27243.
This issue happens when a block with block styles is converted to a reusable block. But it only seems to happen if a post with a block is saved, reloaded, and then the block is converted. Strange.
The issue seems to be that the Block Inspector is still rendering the converted block momentarily after it has been removed. This shouldn't happen, because the Block Inspector renders the selected block, and the block selection is updated as part of the process of converting to a reusable block. Most of the block inspector doesn't seem affected by rendering a removed block (seems like it's not being re-rendered), but the block styles do throw an error if the block doesn't exist, perhaps because they do their own
useSelectcall.In the process of fixing this I changed
BlockInspectorto usewithSelect.How has this been tested?
I've added an end to end test, which fails on
master.Manual steps
Expected: The block should be converted.
Types of changes
Bug fix (non-breaking change which fixes an issue)
Checklist: