cli-runtime/printers: optimize table printing for large result sets

Reduce allocations and CPU time in printTable by:

1. Buffering each row into a reusable []byte slice and writing once,
   eliminating ~3 fmt.Fprint calls per cell (tab, value, truncation).

2. Adding a fast path in WriteEscaped that skips the strings.Replacer
   when the string contains no characters needing escaping (>99% of
   cells).

3. Flushing the tabwriter every 100 rows. With RememberWidths enabled,
   the tabwriter retains column widths across flushes, producing
   consistently aligned output while bounding memory from O(rows) to
   O(1).

Benchmark results (AMD Ryzen 9 7950X3D, go1.24):

  name                            old time/op    new time/op    delta
  TablePrinter_10kRows-32          7.71ms         2.57ms        -66.7%
  TablePrinter_100kRows-32         77.4ms         25.6ms        -66.9%
  TablePrinter_500kRows-32          389ms          129ms        -66.9%

  name                            old allocs/op  new allocs/op  delta
  TablePrinter_10kRows-32         110,066        136            -99.88%
  TablePrinter_100kRows-32      1,100,091        136            -99.988%
  TablePrinter_500kRows-32      5,500,110        138            -99.997%

  name                            old bytes/op   new bytes/op   delta
  TablePrinter_10kRows-32        7,134,294      44,868          -99.37%
  TablePrinter_100kRows-32      75,338,014      44,913          -99.94%
  TablePrinter_500kRows-32     411,596,740      45,153          -99.989%

Kubernetes-commit: 50945f3595cc9a4f4ab2820343f2d3c97ce4abaf