Question

Background: I’m developing a custom regex-like syntax for URL filenames. It will work like this:

• User writes a pattern, something like "[a-z][0-9]{0,2}", and passes it as input
• It is parsed by the program and translated into the set of permutations it represents i.e.
  'a', 'a0', 'a00' … 'z99'

These patterns will vary in complexity, basically anything that could appear in a URL filename must be accommodated. The language is either Java or PHP, but examples in any language or abstract/conceptual help is more than welcome.

My questions are:

1. Where to start with the implementation of a “parser” for the above

and less importantly,

2. How to translate parsed complex patterns into strings programmatically

Answer 1

There is a good answer for this here:
SO: /generate-all-permutations-of-text-from-a-regex-pattern-in-c

The crux of the thing is this…define what you really need well and figure out a way to halt once you have what you need and narrow your search range as much as possible because you are flirting with a quickly exploding number of permutations. “anything that could appear in a URL filename must be accommodated.” is not going to cut it. For example, if you limit yourself to English characters and numbers, for a string 6 characters long you are looking at over 2 billion combinations. For each additional character multiply by 36.
If you go with ISO 8859 you get over 274 trillion combinations and Unicode over 745 trillion-trillion combinations.