Regex to match tags in comments.

Question

I want to get a regex which will match a tag in a java comment so I can replace it with a .net comment.

eg I have this:

/**
* Some method description
* 
* @param paramName Parameter description
*         which may span more than 1 line
* @return return value.
* @throws ExceptionName some exception description
*           again may span more than 1 line.
*/

and I want to end up with this:

///
/// Some method description
/// 
/// <param name="paramName"> Parameter description
///         which may span more than 1 line</param>
/// <returns> return value.</returns>
/// <exception cref="ExceptionName"> some exception description
///           again may span more than 1 line.</exception>
///

The part that I'm not sure about is the best way to deal with matching the @tag and potentially the name afterwards and how to match the inner text that will go between the angle brackets.

Any ideas appreciated.

Answer 1

A single regex is not the best way to go at all here, since you are not doing the same thing with @param and @throws vs. @return. Scripting this in Perl or Python would make this task far easier.

That being said, these should get you on your way ... (All these are in Python's syntax) this:

(?i)(?s)(/\*\*.*?\*/)

will get the entire comment. This:

^(/\*\*|\*/|\*)(.*)$

replaced with ///\2 will turn all of your comment stars with .Net comments.

(?i)(?s)(^([^@]+)@param (\b.*\b)(.*?)(?!\* @return)

replaced with

(\1<param name="\2">\3</param>)

should fix the param portion. Likewise,

(?i)(?s)(^([^@]+)@throws (\b.*\b)(.*?)(?!\*/)

replaced with

(\1<exception cref="\2">\3</param>)

should fix the throws portion. Finally,

@return (.*)

replaced with \1 should fix the @return. (Note this one is not set as multi line [The star does not match EOL characters])

This would all need to be plugged into a script (or else run, one after the other in your editor of choice).

Please note that these do not attempt to account for any stray stars in your @pram, @return, @throw portions.