gastownhall/beads · error

failed to get rows affected: %w

Error message

failed to get rows affected: %w

What it means

After the UPDATE, UnclaimIssueInTx calls result.RowsAffected() and wraps any driver error here. The library needs the affected-row count to detect a lost optimistic-lock race (0 rows), so a failure here means the driver could not report it and the unclaim outcome is indeterminate.

Source

Thrown at internal/storage/issueops/unclaim.go:91

	// oldIssue.RowVersion detects a claim that changed hands (or was released,
	// or closed) between that read and this write exactly as precisely as the
	// old `assignee = <actor>` predicate did, without embedding a
	// spelling-sensitive string comparison in SQL. force does not exempt this
	// check: force only widens WHO may unclaim, not whether the row is still
	// the one we read.
	result, err := tx.ExecContext(ctx, fmt.Sprintf(`
		UPDATE %s
		SET assignee = '', status = 'open', updated_at = ?,
		    started_at = NULL, row_lock = ?
		WHERE id = ? AND status IN ('open', 'in_progress') AND row_lock = ?
	`, issueTable), now, freshRowLock(), id, oldIssue.RowVersion)
	if err != nil {
		return fmt.Errorf("failed to unclaim issue: %w", err)
	}

	rowsAffected, err := result.RowsAffected()
	if err != nil {
		return fmt.Errorf("failed to get rows affected: %w", err)
	}

	if rowsAffected == 0 {
		// The pre-checks passed, so a 0-row result means the row changed
		// underneath us: re-read to disambiguate an ownership change from a
		// status change. actorMatches, not verbatim, mirrors the precheck above.
		current, gerr := GetIssueInTx(ctx, tx, id)
		if gerr != nil {
			return fmt.Errorf("failed to unclaim issue %s: no matching row", id)
		}
		if !force && !actorMatches(current.Assignee, actor) {
			return fmt.Errorf("%w: %s is held by %s; coordinate with the holder — pass --force only if their claim is abandoned (crashed agent, expired lease)",
				storage.ErrNotOwner, id, current.Assignee)
		}
		return fmt.Errorf("failed to unclaim issue %s: no matching row", id)
	}

	return finishUnclaimInTx(ctx, tx, eventTable, id, actor, oldIssue)

View on GitHub (pinned to 71377f2769)

Solutions

1. Inspect the wrapped driver error; treat as transient and retry the whole unclaim if connection-related.
2. Use the supported driver behind the storage driver interface rather than a custom backend.
3. Check driver version compatibility and upgrade if RowsAffected is known-buggy.
4. Verify the connection/transaction is still alive at the point of the call.

Example fix

// before
customStore := NewStore(myOddDriver) // driver lacks RowsAffected support
// after
store := NewStore(supportedDoltDriver) // standard driver via storage boundary

if err := store.ReleaseIssue(ctx, id); err != nil {
	if strings.Contains(err.Error(), "rows affected") {
		// driver-level failure: retry with fresh transaction or switch driver
	}
}

When it happens

Trigger: Driver failing on RowsAffected() after the UPDATE: unsupported driver capability, connection reset, or driver bug mid-transaction.

Common situations: Custom/misbehaving storage driver lacking RowsAffected support; connection dropped between UPDATE and RowsAffected; backend not exposing affected-row counts.

Related errors

• ErrTransaction
• open unit of work: %w
• failed to begin transaction: %w
• failed to commit is_blocked repairs: %w
• failed to begin transaction: %w

AI-assisted analysis of gastownhall/beads@71377f2769 (2026-08-30). Data as JSON: /api/errors/3d89623364d6e91e. Report an issue: GitHub.