Thread: Parsing a complicated string

Hello,

the program I am writting receives a very complicated command line. One of the command line parameter itself actually describes some structure.

For example, the parameter may look something like this:

Code:

QUEUE1[1-50,700-1000],QUEUE2,QUEUE3[1-10,50-100]{1-3,5-5=CUAC,CUAD},QUEUE4{1-4=CUAU}

etc.

Or in other words, it consists of name of queues, than there may be a list of substrings (i.e. take only characters 1 to 50 and 700 to 1000), and than there may be a list of filters (i.e. character 1,2,3 and 5 must equal to 'CUAC' or 'CUAD'), etc.

Because of how the program is called, it has to be one string (no spaces). I am looking for an (academic) advice of how to parse such string effectively. For now, I have only implemented the queue names and substrings, and the parsing logic is getting pretty complicated and pain to modify. On the other hand, I would really prefer not creating a "sort-of" compiler with its own grammar.

I can expect that the command line is correct, and i may modify it a little (instead of ',' use '@', add " if required, or something like that...), but it must remain a single string and must be callable from bash.

Originally Posted by adamm

Thanks for the reply, but I know I can pass multiple parameters in a command line. The problem is, that the string itself is generated by a black-box tool that does not handle spaces correctly

Are you having trouble with passing the string as one parameter or parsing it in the program or both?
What are the rules for parsing the string i.e what information do you need to extract from the string?
Pass it as a single parameter to your program by enclosing that complex string in quotes (single or double).

Dino: I did not consider divide and conquer. Thanks very much, looks viable.