utf8len utf8sub

utf8.lua

function string.utf8len(s)
  if s == nil or s == "" then return 0 end
  local n = 0
  for i = 1, #s do
    local c = string.byte(s, i)
    if c < 0x80 or c >= 0xC0 then
      n = n + 1
    end
  end
  return n
end

function string.utf8sub(s, i, j)
  if s == nil or s == "" then return s end
  if not j then j = #s end

  local bi, ei = nil, nil
  local n = 0
  for ii = 1, #s do
    local c = string.byte(s, ii)
    if c < 0x80 or c >= 0xC0 then
      n = n + 1
    end
    if n == i and not bi then bi = ii end
    if n == j + 1 and not ei then ei = ii - 1 end
  end
  if not ei then ei = #s end
  return string.sub(s, bi, ei)
end