Losing the plot on C# Regex

[DimPrawn]

Hi peeps,

I wonder if any of you .NET guys could try and figure out why my regex works in several tools, but when put into C# it fails?

string    input = @"<a href=""http://www.thisisthedomain.com/"" class='big one' title=""testing & fixing & failing"">text here - More Text - some more</a>";

// Parse HTML for other attributes
Regex regex = new Regex(@"(<a\s+)*title\s*=\s*(?:""|')(?<title>[^'""]*)(?:""|')|class\s*=\s*(?:""|')(?<class>[^'""]*)(?:""|')|href\s*=\s*(?:""|')(?<href>[^'""]*)(?:""|')|rel\s*=\s*(?:""|')(?<rel>[^'""]*)(?:""|')", RegexOptions.IgnoreCase);

Match match = _regex.Match(input);

// Fetch named groups
_rel = match.Groups["rel"].Value;
_href = match.Groups["href"].Value;
_title = match.Groups["title"].Value;
_class = match.Groups["class"].Value; 

Using RegExBuddy, Rad software Regular Expression Designer, Sells Brothers RegExDesigner.NET I can match the "title" group, but my C# fails to (although it gets the "href".

Using 1.1 of the Framework.

If anyone can see what's going wrong I'd be really chuffed.

Cheers,

DP

[scotspine]

this works

Regex regex = new Regex(@"(<a\s+)*title\s*=\s*(?:""|')(?<title>[^'""]*)(?:""|')", RegexOptions.IgnoreCase);

[DimPrawn]

I'm trying to match and capture all the main link attributes, href, title, rel, class in any order in any combination along with any other valid HTML attribute.

Any ideas?

[DimPrawn]

Solved - I think!

Looks like I might have it by doing it in two chunks.

Find all links in HTML:

_regex = new Regex(@"(?<linkhtml><a[\s]+[^>]*?href[\s]?=[\s\""\']+(?<href>.*?)[\""\']+.*?>)(?<linktext>[^<]+|.*?)?<\/a>", RegexOptions.IgnoreCase | RegexOptions.Compiled |RegexOptions.Multiline); 

Then stick the linkhtml group into a 2nd regex:

_regxAttributes = new Regex("(?<name>\\b\\w+\\b)\\s*=\\s*(\"(?<value>[^\"]*)\"|'(?<value>[^']*)'|(?<value>\" + [^\"\"'<> \\s]+)\\s*)+",
                RegexOptions.Singleline | RegexOptions.IgnoreCase | RegexOptions.Compiled); 

To get all the attributes as name and value pairs.

If anyone knows how to capture every hyperlink in the HTML and capture all the attributes in one step I'd be pleased to know.

Cheers,

DP

[scotspine]

that's what i was suggesting. mercifully, life is far too short to spend time on the black art of reg exps...

[TheMonkey]

that's what i was suggesting. mercifully, life is far too short to spend time on the black art of reg exps...

Indeed - which is a good thing because I get paid a lot to maintain them after I've written them

[Joe Black]

that's what i was suggesting. mercifully, life is far too short to spend time on the black art of reg exps...

Indeed, and I'll second that, regex is definately a black art which you'd have to be a C programmer to fully understand.

[TheMonkey]

Indeed, and I'll second that, regex is definately a black art which you'd have to be a C programmer to fully understand.

I'm a C guy too! I think there may be a pattern you have discovered. Something to do with obfuscatable languages... Hmm I'm also a Perl guy. This gets deeper.

[DaveB]

Indeed, and I'll second that, regex is definately a black art which you'd have to be a C programmer to fully understand.

Or a Perl Programmer, for when you really want to confuse people

[TheMonkey]

Or a Perl Programmer, for when you really want to confuse people

Or postscript as well. The following program runs in a postscript and perl interpreter:

http://perl.plover.com/obfuscated/bestever.pl

Insane!