I'm using C# 2.0 on Visual Studio 2005.

I need a regular expression to parse a string with multiple filenames, filepaths, directories and whatnot and separate them. Filepaths often have spaces, so quotes are used to delimit them. There are some invalid characters in a file path, but for this particular application, there's no need to look for them.

For example, when input :
"c:\mes documents\tmp\cator\*.castor" "c:\www\" "file.castor"

There are 3 matches :

c:\mes documents\tmp\cator\*.castor
c:\www\
file.castor

input will always be single line.

I use this regular expression : "\"([^\"]*)\"" and it partially works : I'm getting all 3 matches, but with quotes before and after each match.

In addition, I'd like to have the regular expression also capture the filename if there were no quotes, in that case, each path would be space separated. In fact, filepath validation is not even part of the problem : we can always suppose that the entered paths are valid, and if not, that's a problem I'll solve programmatically ("bad user, bad user!").

Could you help me with such a regular expression?

Please provide an example of your input text that would appear without quotes and specify if such entries may contain spaces in the paths entered.

No, there would be no spaces if there were no quotes

this first example would have the same result as before : i removed the quotes where possible

"c:\mes documents\tmp\cator\*.castor" c:\www\ file.castor

 This example should have 4 matches. Actually, the user makes a mistake, forgetting the quotes, but that's not a problem I want to solve :

c:\mes documents\tmp\cator\*.castor c:\www\ file.castor

matches : 

c:\mes
documents\tmp\cator\*.castor
c:\www\
file.castor 

After thought, the fact that the strings are filepaths is not even part of the problem. Sometimes the paths are absolute, sometimes they are relative, sometimes there are wildcards (there are 2, they are * (asterisk) and ? (question mark) )

You could use something like:

".*?"|\S+

Then strip leading/ending quotation marks from any matches.

I did as you suggested and it worked. Thank you very much.

For reference, here's the final code I used :

public Queue SplitIntoFilePaths (string filepaths) {
	Regex regex = new Regex( "\".*?\"|\\S+" );
	MatchCollection matches = regex.Matches( filepaths );
	Queue q = new Queue( matches.Count );
	string val;
	foreach ( Match m in matches ) {
		string val = m.Value;
		if ( val.StartsWith( "\"" ) && val.EndsWith( "\"" ) )
			val = val.Substring( 1, val.Length - 2 );
		q.Enqueue( val );
	}
	return q;
}