Devana fixes - #9
Conversation
Pass fallbackLayout to the row LayoutPatternField so blurring an empty or invalid layout draft normalizes back to the current layout instead of "", which previously truncated multi-column rows to the first column on blur. Resolves devana finding layout-blur-drops-columns.
…on-row-update) LayoutPatternField unconditionally re-synced its draft from the value prop, so same-row structural edits that change row.layout could wipe uncommitted text. Add an isFocused ref and skip the sync effect while focused; clear it on blur before committing so the committed value still syncs back. Resolves devana finding layout-draft-lost-on-row-update.
…-value-mishandled) A persisted layout value that is a single row object (instead of an array) was silently treated as empty in admin and threw in normalizeLayoutRows on render. Add a shared asLayoutRows helper that wraps a recognized singleton row into a one-element array; route both admin asLayouts and render ingest through the same isLayoutBuilderRow check. Non-row objects still map to []. Resolves devana finding non-array-layout-value-mishandled.
…ovable) The layout menu (incl. the remove option) was only built when layouts.length > 1, so a single row could never be removed back to the empty [] state. Drop the outer gate; move-up/down already self-gate by row position, so only the remove option shows for a lone row. Resolves devana finding last-layout-row-not-removable.
…ids-unstable) normalizeRow fabricated layout-implied columns with randomId on every render, so an unrelated parent re-render before the first save reassigned ids and remounted their block editors (losing in-progress edits). Switch the factory to deterministic layout-N-column-M ids, matching normalizeColumn and render's normalizeLayoutRow. randomId stays for user-initiated add actions. Resolves devana finding synthetic-column-ids-unstable.
…yout-row-null-throws) normalizeLayoutRow read layout.columns directly, so a null/undefined/primitive hole in the layouts array threw a TypeError and aborted the Astro render path while the admin path tolerated it. Coerce the argument via isRecord before any property access; holes now synthesize a default row, matching admin. Resolves devana finding normalize-layout-row-null-throws.
…-row-span-gap) isLayoutBuilderRow only inspected columns[0] for a span, so legacy rows carrying span on a later column (with no top-level layout key) were misclassified as non-layout even though normalizeLayoutRow accepts them. Scan every column with an isRecord guard, which also makes the in-check safe on primitive column entries. Resolves devana finding is-layout-builder-row-span-gap.
…umn-blocks-mismatch) A column whose blocks was a singleton object (not an array) was silently emptied in admin and threw in render (package normalizeBlocks maps over it). Add a shared asBlocksArray helper that wraps a recognized singleton block into a one-element array; route both admin normalizeBlocks and render's normalizeLayoutColumn through it. null/primitive blocks still map to []. Resolves devana finding non-array-column-blocks-mismatch.
…ize-layout-column-null-throws) normalizeLayoutColumn dereferenced column.id directly, so a null/primitive hole in the columns array threw and aborted render while admin tolerated it. Coerce the column via isRecord before any property access; holes now synthesize a default column, matching admin. Resolves devana finding normalize-layout-column-null-throws.
…g-false-wrap) Per-span Math.round overflowed the 12-column grid for equal-width patterns whose fractions sum to 1/1 (seven 1/7 rounded to 2 each = 14 > 12), wrapping rows the README says should fit on one line. Add layoutGridSpans: for rows summing to <=1/1 it distributes round(total*12) units via largest-remainder (each >=1); rows summing to >1/1 keep per-span rounding so they still wrap. Admin uses one rowGridSpans per row; helper is re-exported for frontends. Resolves devana finding span-rounding-false-wrap.
…builder-row-primitive-throws)
The TypeError on 'span' in <primitive> was already eliminated by the
span-gap rewrite (isRecord guard per column). Add the exact counterexamples
({ columns: ["x"] }, [0], [true], ["a","b"]) as explicit non-throwing/false
assertions and mark the finding fixed.
Resolves devana finding is-layout-builder-row-primitive-throws.
…age-default-shadows-fallback) bentoMessage short-circuited to the built-in en default the moment the default locale appeared in the fallback chain, so a chain routing fr->en->de returned "Grid" instead of the de override "Raster" — diverging from localizedString. Remove the in-loop default-locale short-circuit; walk the full chain for user overrides and resolve the built-in default only after the loop. Resolves devana finding bento-message-default-shadows-fallback.
…-key-id-collision-state-bleed)
A stored id could equal a sibling's synthesized positional id (e.g. a row
id'd "layout-2" next to an id-less row at index 1 -> layout-2), producing
duplicate key={row.id}/key={column.id} values and bleeding per-instance editor
state across cards. Add a deterministic uniqueId pass in asLayouts: dedupe row
ids globally and column ids per row, deriving a stable id on collision. Stays
deterministic (not randomId) so ids remain stable across re-renders.
Resolves devana finding react-key-id-collision-state-bleed.
|
Warning Review limit reached
Next review available in: 42 minutes Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available. How can I continue?After more reviews become available, a review can be triggered using the To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews. How do review limits work?CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability. For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window. Please refer docs for additional details. Review details⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: ⛔ Files ignored due to path filters (1)
📒 Files selected for processing (11)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: ac99170db3
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| const fractions = spans.map(spanFraction); | ||
| const totalFraction = fractions.reduce((sum, fraction) => sum + fraction, 0); | ||
|
|
||
| if (totalFraction > 1) { |
There was a problem hiding this comment.
For a valid full row like eleven 1/11 spans, JavaScript sums the fractions to 1.0000000000000002, so this strict > 1 check takes the overflow path and returns eleven 1-column widths instead of distributing the row across all 12 grid columns. That leaves an unused grid column in the admin preview and in frontend renderers using the new README helper; compare with an epsilon or exact rational total before deciding the row is wider than one line.
Useful? React with 👍 / 👎.
1 participant
No description provided.