Help decyphering EVIL RegEx's

This is why I hate RegEx's so much. Can someone explain the conditions for the 2 RegEx's here: I'll much appreciate it.

Regex aclQueryStringRegEx = new Regex(".*[&?]acl($|=|&).*");

Regex torrentQueryStringRegEx = new Regex(".*[&?]torrent($|=|&).*");

if (aclQueryStringRegEx.IsMatch(resource))

{

   sb.Append("?acl");

}

else if (torrentQueryStringRegEx.IsMatch(resource))

{

   sb.Append("?torrent");

}

littleguru wrote:

Don't they just say: if you encounter a list of chars then &torrent= or &acl= we have a match?

What are you doing with torrents, you h4x0r?

You mean that could be expressed as:

if (resource.IndexOf("?acl") >= 0)

   sb.Append("?acl");

else if (resource.IndexOf("?torrent") >= 0)

   sb.Append("?torrent");

If so, I just don't understand people at times. And, no. I'm no l33t h4x0r

I'm playing around w/ the Amazon S3 (storage) service. Pretty cool so far

Minh wrote:


You mean that could be expressed as:

if (resource.IndexOf("?acl") >= 0)

   sb.Append("?acl");

else if (resource.IndexOf("?torrent") >= 0)

   sb.Append("?torrent");

More like this one:

if ((resource.IndexOf("?acl") >= 0) || (resource.IndexOf("&acl") >= 0)
)

   sb.Append("?acl");

else if ((resource.IndexOf("?torrent") >= 0) || (resource.IndexOf("&torrent") >= 0)
)

   sb.Append("?torrent");

If you need help decyphering regular expressions, try Expresso http://www.ultrapico.com/Expresso.htm; you can paste existing expressions into the application and get it to analyse them for you, explaining each step.

To be honest, this particular example does seem like overkill but on the whole regular expressions are immensely valuable and are worth your time and effort learning.

Minh wrote:

This is why I hate RegEx's so much. Can someone explain the conditions for the 2 RegEx's here: I'll much appreciate it.

Regex aclQueryStringRegEx = new Regex(".*[&?]acl($|=|&).*");

Regex torrentQueryStringRegEx = new Regex(".*[&?]torrent($|=|&).*");

if (aclQueryStringRegEx.IsMatch(resource))

{

   sb.Append("?acl");

}

else if (torrentQueryStringRegEx.IsMatch(resource))

{

   sb.Append("?torrent");

}

Matches any string that contains &acl (&torrent in second) or ?acl (?torrent) somewhere other than the beginning of the string, optionally followed by other arguments, the end of the string or an =... parameter.

So basically it is asking if the thing (which is probably, but not definitely and URL) contains a GET parameter torrent or acl respectively.

If it contains a get ACL parameter, sticks a ?acl to the end of sb, xor it contains a TORRENT parameter, sticks a ?torrent to the end of sb.

Thanks, guys. I missed the &acl and &torrent bits. But what's the deal with the = sign?
And I'll give Expresso a shot too. By the sheer existance of Expresso, doesn't that tell us all that RegEx's are bad?

Minh wrote:

Thanks, guys. I missed the &acl and &torrent bits. But what's the deal with the = sign?
And I'll give Expresso a shot too. By the sheer existance of Expresso, doesn't that tell us all that RegEx's are bad?

No it doesn't - in exactly the same way that the existance of Visual Studio doesn't mean that C# is evil...

The ($|=|&) matches either:
* end of line or string
* equals character
* ampersand character

I think I'd prefer the RegEx as:

if (CheckForCommand(resource, "?acl") || CheckForCommand(resource, "&acl"))

   sb.Append("?acl");

else if (CheckForCommand("?torrent") || CheckForCommand(resource, "&torrent"))

   sb.Append("?torrent");

---

private bool CheckForCommand(string resource, string command)
{
   int pos = resource.IndexOf(command);
   if (pos < 0) return false;

   int posEnd = pos + command.Length;
   if (posEnd == resource.Length) return true;

   char post = resource[posEnd + 1];
   if (post == '=' || post == '&')
      return true;
   else
      return false;
}

IceFreak2000 wrote:



No it doesn't - in exactly the same way that the existance of Visual Studio doesn't mean that C# is evil...

Well, the alternative for C# would be to write MSIL,

but the alternative to RegEx is code. And I'll take code over RegEx any day.

http://regexlib.com/RETester.aspx

Minh. though this seems like very technical questions you ask, i am glad you asked it here in the coffee house. i just learned so much. and it blows my mind away how smart you and so many other people can be. humbling i think is the word. i had to ask wiki, and a friend of mine for some help on learning what regular expressions are and how they work.

I dont know why you choose to place the characteristic of EVIL on Regular Expressions. They dont seem to be Evil nor Good. so far...

funny thing is, we seem to use something like regular expressions in our every day lives when we comunicate but if asked to explain them, this would be my best start.

Their evil because their generally slow, and most people will struggle to understand a regular expression..

I know I've written some in the past.. come back to them later on.. looked at them and not exactly sure what they were doing.. thank god I've always been 'comment / xml comment' 'happy'.

Yeah. What stevo_ said.

The intent of the code cannot be transmitted to people who don't know the RegEx syntax, but could be lost to the original author, who might have to deal w/ RegEx's that one time 6 months ago.

Minh wrote:



IceFreak2000 wrote:  No it doesn't - in exactly the same way that the existance of Visual Studio doesn't mean that C# is evil...

Well, the alternative for C# would be to write MSIL,

but the alternative to RegEx is code. And I'll take code over RegEx any day.

What I meant was that you can write C# code using Notepad and compiling from a command prompt, or you can use a full-blown IDE with IntelliSense support such as Visual Studio.

In the same way, you can easily write regular expressions 'by hand', but some people find that utilites like Expresso a great help - especially when they're not overly familiar with regular expression syntax.

Minh wrote:

Yeah. What stevo_ said.

The intent of the code cannot be transmitted to people who don't know the RegEx syntax, but could be lost to the original author, who might have to deal w/ RegEx's that one time 6 months ago.

Actually, they're not that slow at all - remember that if you declare a regular expression with the RegexOptions.Compiled option the expression is compiled to an assembly. While this adds a small amount of overhead on first use, any subsequent use is very fast indeed.

As an example, the following regular expression will match all repeated words such as in

Paris in the
the spring

Regex regex = new Regex(@\b(?<word>\w+)\s+(\k<word>)\b",
  RegexOptions.Compiled | RegexOptions.IgnoreCase);

Now think about how much harder it would be to do the same thing without regular expressions.