utf8 codepoint regex

Textadept uses TRE as its regex engine for text searches. Unfortunately, TRE does not allow for easy searching of UTF-8 codepoints (e.g. U+2318). Normally, UTF-8 codepoints would have to be broken up into their individual bytes and searched for separately (e.g. \xe2\x8c\x98). Adding the following to your ~/.textadept/init.lua extends regex syntax to silently convert "\u{codepoint}" sequences into the individual byte sequences needed by TRE. Thus a search like "\u{2318}" would work as expected.

local overriding_find = false
events.connect(events.FIND, function(text, next)
  if not overriding_find and ui.find.regex then
    -- Substitute \u{codepoint} with \xXY byte sequences needed by TRE.
    text = text:gsub('\\u{(%x+)}', function(codepoint)
      local char = utf8.char(tonumber(codepoint, 16))
      return string.rep('\\x%X', #char):format(char:byte(1, #char))
    end)
    overriding_find = true
    events.emit(events.FIND, text, next)
    overriding_find = false
    return true -- prevent normal 'find' handling
  end
end, 1)