[mp4]

Hi!

I found some bugs again and checked in the new version.
I have not introduced this library yet, I will do so now.

I wanted it to be Perl compatible so it is almost compatible.(what is implemented)
Current weakness or it could be faster:
-for verbosely defined character classes, it is not as efficient as it could be..
E.g. [abcdopzxcvpiuy]
-regular expressions that start only with ^ or end with only $
-for complicated regular expressions with many choices (|). This is a property
of backtracking algorithms. All similar designs have the same problem. (e.g Perl regular engine). Execution times can be exponential.

Strengths:
-the regular expressions are converted into code at compile time.
-it can handle some expression that others don't. I have just stumbled into this example (matching C-like strings): \"(?:\\\\|\\\"|\\\n|[^\n])*?\".
This crashed with Greta in fast mode and in Pcre 4.2 for long 8000 byte strings
-better more simple interface . This may be subjective but it was a goal.