* removed 100 limit on file search results because we
  load all results anyway
* expensive processing (parsing match content, utf encoding)
  is done only for selected page in paginated output