正则表达需要很长时间才能完成

Question

I wanted to match a line that is structured like this:

• Start of line
• Multiple '-'
• Maybe a white space (maybe more)
• At least one character
• Maybe more characters and white spaces
• Maybe a white space (maybe more)
• Multiple '-'
• End of line

So I wrote the Regex like this:

new Regex(@"^\-{2,}\s*(\w+(\w+|\s)*)\s*\-{2,}$");

And when I try to match the following line, this takes ages to complete (didn't wait for it to complete):

-------- Variable used for recipe visualization only - Not loaded into PLC --------

I think there's a very big number of matches in it and the Regex have hard time enumerating all those matches but I'm not sure.

Environment information: Windows 7, framework 3.5

Thank you

Edit: Thanks to your help I came up with a Regex that works:

^-{2,}\s*(?!\-)(\w(?:\w|\s|\-)+)(?<!\-)\s*-{2,}$

So the interpretation:

• Start of line
• At least two '-'
• Maybe a white space (maybe more)
• No more '-'
• At least one character
• Maybe more characters, white spaces or '-'
• No more '-'
• Maybe a white space (maybe more)
• At least two '-'
• End of line

If you see something wrong with it please tell me

Answer 1

Unroll the nested grouping as

^-{2,}\s*(\w+(?:\s+\w+)*)\s*-{2,}$
             ^^^^^^^^^^^ 

Otherwise, your pattern will be prone to catastrophic backtracking .

See the regex demo

Alternatively, use an atomic group to disable any backtracking into the alternation group:

^-{2,}\s*((?>\w+(?:\w+|\s)*))\s*-{2,}$
          ^^^              ^ 

See this regex demo

Generally, avoid alternations with nested quantifiers (like in (\\w+|\\s)* ) inside longer patterns.