npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

@bsuite/data-grid

v3.0.5

Published

Virtualized, spreadsheet-feel data grid for BSuite apps — cell range selection, keyboard grid navigation, TSV clipboard round-trip, fill handle, undo/redo. Built on @tanstack/react-table + @tanstack/react-virtual.

Readme

@bsuite/data-grid

Partial persistence outcomes

onCellsEdited still accepts synchronous or asynchronous void (every cell succeeded). A host that saves a batch independently can return CellEditResult:

<DataGrid
  data={rows}
  columns={columns}
  getRowId={(row) => row.id}
  onCellsEdited={async (edits) => {
    const failures: CellEditFailure[] = [];
    for (const edit of edits) {
      try {
        await saveCell(edit.rowId!, edit.columnId, edit.value);
      } catch (error) {
        retainRetryDraft(edit); // Host-owned draft, separate from the grid's data.
        failures.push({
          rowId: edit.rowId!,
          columnId: edit.columnId,
          status: 'failed',
          message: error instanceof Error ? error.message : undefined,
        });
      }
    }
    return { failures };
  }}
  onError={showSaveError}
/>

Import CellEditFailure and CellEditResult from @bsuite/data-grid. Every edit emitted by the grid includes rowId; the field remains optional in CellEdit for compatibility with existing host-created objects. Match outcomes using rowId and columnId, never the current display index. getRowId must return unique, stable record identifiers when rows can be reordered or filtered. The legacy index fallback cannot identify records across a host data reorder.

Each omitted cell is successful. Failures require status: 'failed'; message is optional and defaults to “Failed to save the edit.” Unknown or duplicate identities invalidate the result and report the submitted cells as failed; the host must refresh before retrying an uncertain write. Rejecting the callback continues to mean the whole batch failed. A partial saver must therefore return the actual failed subset instead of throwing after some writes have succeeded. Each applyCells (or retry) call resolves with the receipt for that exact invocation, even if a newer attempt later supersedes one or more of its cells.

Failed optimistic values are removed and the corresponding cells show their refusal message. Host data/refetch controls their displayed value; discarding a host retry draft cannot leave a hidden grid overlay behind. Successful overlays remain until host data acknowledges that value, then release so later server updates are visible. The host must update or invalidate its authoritative data on persistence. A fresh edit clears that cell's old refusal.

Undo contains only successful cells. Partial undo and redo move successful cells to the opposite history and retain refused cells on their original side for retry. An unavailable row is refused locally instead of sending a write to the record now occupying its former index. A pending save can be undone; its late completion cannot replace the newer overlay or undo history.

Per-cell ownership protects grid state, not database write ordering. Hosts must serialize same-cell writes (including undo/redo) or use server conditional writes. A UI ownership check cannot cancel an already-running database update. When a host serializes whole batches, capture its promise queue outside render, chain each batch onto it, and recover the queue's rejection without converting a failed batch into success. Keep latest-attempt ownership for host draft/error state as well. onError receives only currently owned failed cells; it does not report stale failures belonging to an older attempt.

Retrying a retained draft

Use gridRef.current.applyCells([{ rowId, columnId, value }]) with the already typed values from failed CellEdits. It rebinds each identity to the current row and uses the same refusal, persistence and undo pipeline as direct editing. The promise settles after persistence outcomes are handled; inspect the host draft store for any remaining failures. Missing/duplicate identities reject before any write. Do not call the host persistence callback directly for retry: that would leave grid refusal and history state disconnected from the save.

All value mutations, including paste/fill/clear/undo/retry, require an explicitly editable column. Linked fields must use their supplied record picker and the link mutation callback; a pasted label is never a record reference.

Implementation and verification bounds

This extends the existing grid and undo stack; no replacement engine or new dependency is used. Stable identity follows the installed TanStack Table 9.2.4 table_getRowId implementation (custom extractor, otherwise string index), confirmed with Context7 and the installed package's dist/core/rows/coreRowsFeature.utils.js. See the official TanStack row identity guide. The alternative of retaining failed optimistic drafts inside the grid loses refetch/discard authority and was rejected; retry editing belongs to the host.

Mounted tests exercise actual grid keyboard paste, mixed failures, refetch, undo/redo, reorder/filter, no-op/void behavior, and overlapping completions. ui_touched=true: the consumer rollout remains open across the 23 existing CRM render sites, the new Conduit Candidates surface, and other estate consumers. Package tests do not establish deployed UX acceptance. Consumer wiring, preview publication/pinning, and the parent's D8/live verification remain required.

Per-cell edit restrictions

Stable 3.0.5 finalises the published 3.0.5-next.0 source without a runtime change. Development consumers use the preview artifact for deployed UX testing; stable publication and consumer adoption remain separate release steps. This parent source integration does not update any app lockfile. Follow the preview publishing standard.

Use isCellEditable={(row, columnId) => !row.imported} for authoritative rows that must remain read-only inside an otherwise editable grid. This optional, pure, synchronous predicate only restricts editing: columns still require editable: true, and linked fields still require their existing record picker. Omitting it preserves column-only editability.

The predicate receives the latest host row, without optimistic values. It controls mouse/keyboard editor entry and aria-readonly, paste/clear/fill targets, and the mutation boundary for retry/undo/redo. Protected targets are skipped before parsing or optimistic display during paste/fill/clear; TSV positions are preserved, so values never shift into a different row. Explicit applyCells and history operations return normal per-cell refusal receipts for currently protected cells. Eligible manual siblings in a mixed batch still save. Selection, navigation and copy remain available; a wholly read-only selection has no fill handle. Mixed selections retain the handle, but protected destination cells remain unchanged.

Open editors follow stable row/column identity and visual coordinates through reorder; permissions are checked again at commit. If their row or column disappears, the grid cancels the invisible edit, reports that no edit was saved through onError, and restores keyboard input. Queued undo rechecks current policy after waiting for its predecessor. Hosts must retain their own server and mutation guards: this API prevents impossible UI edits and does not provide database authorization.

The alternative of disabling an entire grid would prevent work on manual rows; consumer event interception would duplicate mutation rules across mouse, keyboard, clipboard and history. One shared predicate keeps those paths consistent. The mounted regression covers both frozen/scrollable cells, linked pickers, mixed mutation ranges and policy changes during active editor/clipboard/history work. Accessibility follows the WAI-ARIA grid pattern.