perf: a batch of small optimizations

[lukecotter]

I've been profiling Tabulator on large datasets and ended up with a handful of small, self contained performance PRs.
No behaviour chnage on any, but they're related so I figured a short overview in one place would help.

The PRs:

• perf(core): clone row template and slice active rows #4904 – clone the row template and use slice() for the active-rows copy — grid filter −23%; setActiveRows 166× isolated
• perf(events): avoid arguments/forEach allocation in event buses #4905 – avoid arguments/forEach allocation in the event buses - grid render −36%; dispatch 7–9× isolated
• perf(sort): precompute sort keys and fast-path numeric sorter #4906 – precompute sort keys and fast-path the number sorter — grid sort −31%; _sortItems 2.6× / number sorter 4.9× isolated
• perf(modules): reduce closure allocation in filter and calc loops #4907 – reduce closure allocation in the filter and calc loops — filterRow up to 3.8×, generateRowData 2× isolated (negligible in a full grid)
• perf(datatree): cache module refs and flatten child recursion #4908 – cache module refs and flatten the data-tree child recursion — getChildren 1.6× isolated (negligible in a full grid)
• perf(frozencolumns): reinitialize rows in O(n) #4909 – reinitialize frozen-column rows in O(n) instead of O(n²) — reinitializeRows 1.5× isolated (negligible in a full grid)
• perf: streamline cell content generation #4724 – drop a redundant innerHTML write in cell content generation — render 2× / clear 2.6× isolated (negligible in a full grid)
• perf: track deepClone circular refs with a WeakMap #4734 – track deepClone circular refs with a WeakMap (also fixes a circular ref clone bug) — 38–147× isolated (off the render path)

Everything was measured against master (9539446). A few of these have a noticeably perf impact.
With all of them applied, a typical render → sort → filter pass over 20k rows went from ~56ms to ~30ms (roughly 1.8× faster), mostly from the render and sort paths.

[lukecotter]

I also have some renderer fixes. All those Prs are listed on #4938.
kept separate from this one because those PRs are mainly fixes not perf.