CODEWITHSUNDEEP
rubyregex
user1759065
user1759065

Reputation: 1

ruby regex for php comment block

I've been trying to find the regex in ruby to match a php comment block:

/**
 * @file
 * lorum ipsum
 * 
 * @author  ME <me@localhost>
 * @version 00:00 00-00-0000
 */

Could anyone help I've tried searching alot and even though some regex I found has worked in a regex tester but doesn't when I write it in my ruby file.

This is the most successful bit of regex I have found:

 (/\*([^*]|[\r\n]|(\*+([^*/]|[\r\n])))*\*+/)

This is the output from my script 

file is ./test/123.rb so regex is ((^\s*#\s)+(.*?))+
i = 0
found: my first ruby comment
file is ./test/abc.php so regex is (/\*([^*]|[\r\n]|(\*+([^*/]|[\r\n])))*\*+/)
i = 0
found: * 
i = 1
found: *

Here is the code I have to do this:

 56   def self.extract_comments f
 57     if @regex[File.extname(f)]
 58       puts "file is " + f + " so regex is " + @regex[File.extname(f)]
 59       cur_rgx = Regexp.new @regex[File.extname(f)]
 60       matches = IO.read( f ).scan( cur_rgx )
 61       content = ""
 62       if ! matches.empty?
 63         # content = "== " + f + " ==\n"
 64         content += f + "\n"
 65         for i in 0...f.length
 66           content += "="
 67         end
 68         content += "\n"
 69         for i in 0...matches.length
 70           puts "i = " + i.to_s
 71           puts "found: " + matches[i][2].to_s
 72           content << matches[i][2].to_s + "\n"
 73         end
 74         content << "\n"
 75       end
 76     end
 77     content || '' # return something
 78   end

Upvotes: 0

Views: 165

Answers (2)

skunkfrukt
skunkfrukt

Reputation: 1570

Unless it is important that each line inside the comment block begins with an asterisk, you may want to try this regex:

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

EDIT: And here's a stricter version, which will only match comments that are formatted exactly like your example:

/^( *)\/\*\*\n(?:\1 \*(?:[^*\n]|\*(?!\/))*\n)+\1 \*\//

This version will only match a comment that has /** and */ on separate lines. /** can be indented by an arbitrary number of spaces (but no other white-space characters), but the other lines must be indented by exactly one more space than the /** line.

EDIT 2: Here is another version:

/^([ \t]*)\/\*\*.*?\n(?:^\1 .*?\n)+^\1 \*\//

It allows a mixture of tabs and spaces (ew) for indentation, but still requires all lines to conform to the indentation of the /** one (plus a single space).

Upvotes: 0

pguardiario
pguardiario

Reputation: 55002

It seems like /\/\*.*?\*\//m should do. Also that's really a c-style comment block.

Upvotes: 1

Related Questions