What is this following regular expression saying?

Just came across this following regular expression:

Regex.Match(feed.Element("description").Value, @"^.{1,180}\b(?<!\s)").Value
开发者_StackOverflow

I know it says that starting with anything it should contain minimum 1 and maximum 180 characters \b stands for word boundaries. I didn't understand what is \b doing here. And then (?<!\s). What is that expression doing? ?<! stands for look behind and don't consume string. My guess is that it says look behind and it shouldn't end with space. I am not sure though. Can anybody clear these doubts.

---

See your expression here on Regexr, this is a useful tool to test regular expressions.

I reduced the maximum length to 10 for test. So it looks like

^.{1,10}\b(?<!\s)

(?<!\s) is a negative look behind zero length assertion. That means it checks if the position before (on the left) is not a whitespace.

So, ^.{1,10}\b(?<!\s) will match at the last word boundary in the first 10 characters of the string, but only if the left part or the word boundary is not a whitespace. This will not only match on "left word boundaries" (I think tripleee means the right side of the word), because word boundaries does not necessarily include whitespace.

A word boundary \b will match on the empty string between a word character (defined by the class \w) and a non word character \W.

That means \b(?<!\s) will match for example between "A$", "A ", "(A" or ".A". All of them have a word boundary in between and the left character is not a whitespace.

---

In your case (?<!\s) makes sure that trailing whitespaces are not included in the match

It's easy to illustrate the following way. Change 180 to 10 in your example, so you do not need a really long test string:

^.{1,10}\b(?<!\s)

Now try to match the following string against it (note two spaces between two and three):

one two  three four

Your regular expression match won't include the two whitespaces between two and three. However if you remove the last part of your regular expression like this:

^.{1,10}\b

Then the two spaces after two will be included in the match.

---

Basically, the "not whitespace" assertion forces the \b to match left word boundaries only. So in other words, if the first 180 characters contain a word start anywhere after the first column, this matches. (The expression requires at least one arbitrary character before the match -- hard to say without context whether this is really correct, and what exactly it is supposed to accomplish.)