Regex matching javadoc comment for a specific method (python)

Question

I'm having trouble coming up with a regex that will match the javadoc comment contents for a specific java method. Example:

/**
 * Do not match this.
 */

/**
 * Do match this.
 */
@SomeAnnotation
public boolean methodX() { }
/**
 * Do not match this.
 */

I already know the method signature so I can use that in the regex.

I can match all of the javadoc comments using:

/\*\*(.*?)\*/

I'm also specifying re.DOTALL. I tried expanding the regex to use a negative lookahead that says I only want a javadoc comment if it's the comment immediately proceeding the method:

/\*\*(.*?)\*/(?!.*?/\*\*.*?public boolean methodX\(\))

But that's causing the (.*?) to match the contents from the start of the first javadoc comment to the end of the javadoc comment immediately proceeding methodX.

I keep trying various ways of constructing positive and negative lookaheads but nothing is working. What am I missing?

Are you looking to capture the annotation as well? And are you trying to capture and use the text portion of the comment that you match? — wolffer-east
– wolffer-east, Commented Jul 10, 2014 at 18:50
I just need the text portion of the comment, but I can strip the *'s and extra newlines as a post-processing step. The text is fairly structured so I'm not too worried about weird edge cases. Most of the methods have several annotations so I just wanted to make sure that was communicated in the example. — tutu056
– tutu056, Commented Jul 10, 2014 at 19:54

famousgarkin · Accepted Answer · 2014-07-10 18:48:23Z

2

This matches the comment (from /** to */) preceding the function in the given example text in a comment named group:

(?P<comment>/\*\*(?:(?!/\*\*).)*?\*/)(?:(?:(?!\*/).)*?)(?=public boolean methodX)

See a test at regex101.com.

The key here is to ignore the extra /** and */ in the wanted text using (?!/\*\*).)*? and (?!\*/).)*?
?:s are to scrape the uninteresting groups from the result

edited Jul 10, 2014 at 18:48

answered Jul 10, 2014 at 18:35

famousgarkin

14.2k5 gold badges60 silver badges74 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

wolffer-east Over a year ago

This has some odd behavior during a fringe situation: where an annotation might contain */. This regex grabs the entire annotation up to that point as a part of the comment, when it should not.

famousgarkin Over a year ago

@wolffer-east I see, how does such annotation look like? Could you provide an example please?

wolffer-east Over a year ago

The following would have '@SomeAnnotation that contains ' as part of the comment capture: @SomeAnnotation that contains */ is a troublesome annotation. If you follow the link in famousgarkin's post and add */ to the end of the annotation you will see the capture change.

wolffer-east · Accepted Answer · 2014-07-10 20:22:34Z

1

Your expression is greedy and is currently matching the */ in the first comment (because .* matches */). try using

/\*\*((?:[^*]+|\*[^/])*)\*/

This ensures that you will never match the ending */ by accident and end up with two comments matched at the same time

EDIT: This code avoids the issue of annotations that contain */. not sure why they would, but here goes:

/\*\*((?:(?!\*/).)*)\*/(?:(?!/\*\*).)*(?=public boolean methodX)

check out this example for confirmation that it works: http://regex101.com/r/yV9oK2/2 I switched from my original match to a negative lookahead to avoid a 'catastrophic backtrack' as the test program put it :)

edited Jul 10, 2014 at 20:22

answered Jul 10, 2014 at 18:12

wolffer-east

1,0697 silver badges15 bronze badges

Collectives™ on Stack Overflow

Regex matching javadoc comment for a specific method (python)

2 Answers 2

3 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

3 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related