improve binary parsing of isodat files

[sebkopf]

take advantage of several more specific codes for faster interpretation of results and to avoid slow regular expressions:

• FF FE FF x = always followed by x (byte interpreted as number) number of unicode characters, could just parse the entire file for text before doing anything else, same as we do with the C-blocks)
• x 0 0 0 = structural elements, sometimes data types markers

list of things to implement

• structure analysis (blocks field in binary file)
• rewrite file navigation in terms of the overall blocks rather than just C-blocks
• rewrite peak table parsing based on the structure blocks instead of slow loops
• rewrite file info parsing based on the structure blocks
• rewrite gas config parsing and raw data based on the structure blocks
• rewrite the rest based on the structure blocks for cf files
• rewrite scan files read based on structurte blocks
• rewrite did files read based on structure blocks
• make binary_file class isodat_binary_file class (extends binary_file)
• introduce is_binary_file, is_isodat_binary_file for easier use in stopifnot checks
• rename binary to source throughout
• test with all available files to make sure everything still works as intended