Multiline regex matches get written into output instead of regex match space

[SoptikHa2]

Example of sed debug output:

MATCHED REGEX REGISTERS
  regex[0] = 0-12 ',3;
:1 2:3 4'
  regex[1] = 0-9 ',3;
:1 2:'
  regex[2] = 1-2 '3'

The way it works now is that only the regex[0] line is read. When we parse next line, it looks like output from sed, so it gets written that way. Actually, all next four lines are considered sed output. We need to handle regex that has newline in it. We can probably ditch all the escaping hell with '. Who would ever need it.