You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Blocks field: block added while form state is still syncing renders empty (header only, no fields) #18638
Payload version: 3.90.2 (Next.js 16.3.3, @payloadcms/db-postgres, admin language German)
If a block is added while the debounced form-state request triggered by a previous keystroke is still pending, the new block renders with its header only and without any fields. It stays empty until reload. Reliably reproducible in a production build (next build && next start) when adding a block immediately after typing in another field.
Steps to reproduce:
Open a collection with a blocks field in a production build.
Type into the title field and immediately (within the debounce window) click "Add block" and pick a block.
Actual: the block row appears without fields.
Expected: the new block's fields render regardless of pending form-state requests.
Possibly related: #17375 (pasted block row renders blank until refresh). Probable cause is the form-state response for the earlier request overwriting the state that contains the new row.
How we found it: axe-core (tags wcag2a, wcag2aa, wcag21a, wcag21aa, wcag22aa) via @axe-core/playwright plus keyboard-only tests of the admin. Our tests wait until no request is pending before adding a block.
Happy to help with a PR if the direction is agreed.
Hi, I'd like to take this issue for my open-source contribution sprint. I'll start by reproducing the production-build case and tracing how the pending form-state response interacts with the newly added block.
My proposed scope is a focused synchronization fix with a deterministic regression test that delays the earlier response, adds a block, and checks that its fields become editable without reloading or losing typed values. I'll also check the overlap with #18425 and the earlier pasted-block work before implementing.
Please let me know if someone is already working on this, or if you have a preferred direction or target branch. Thanks!
Activity
dharshni1518 commentedon Oct 10, 2026
Hi, I'd like to take this issue for my open-source contribution sprint. I'll start by reproducing the production-build case and tracing how the pending form-state response interacts with the newly added block.
My proposed scope is a focused synchronization fix with a deterministic regression test that delays the earlier response, adds a block, and checks that its fields become editable without reloading or losing typed values. I'll also check the overlap with #18425 and the earlier pasted-block work before implementing.
Please let me know if someone is already working on this, or if you have a preferred direction or target branch. Thanks!