REGEXP: In-line comments

Question

I am building a formatter for MU* game coding (MUSHes, if you are familiar). In short, the program takes this:
&TEST me=$test:     
  [setq(0,#2)];
## Test Comment
  @switch/first
    [ifelse(1,1,0)]  ## Inline test comment.
  =
  0,
##          Another Comment with more spaces
      @pemit %#=%ch%crERROR:%cn Uhoh!
  1,
      @pemit %#=%ch%cgGAME:%cn Good!
    }parses through it, and produces a single line:
&TEST me=$test:[setq(0,#2)];@switch/first [ifelse(1,1,0)]=0,{@pemit
%#=%ch%crERROR:%cn Uhoh!},1,{@pemit %#=%ch%cgGAME:%cn Good!}NOTE: The first step is pulling all of the comments (This is where I am having trouble), I will work on removing newlines and spaces, et al, after I overcome this hurdle.
The problem I am running into is that the '## COMMENT' regexp I am trying to build is not working. I am currently using \#\#[(\s+)(\S+)]+$, but this is ending up with everything after the [setq()] function being axed. All I want to do (at this first point) is remove everything from '##' to the end of that input line.
In my readings, I am finding that the way Java parses MULTI-LINE strings is very different from other languages, such as Perl. Java wants to match the /entire/ string, which I do not want.
How do I flag the pattern object to say 'Hey, only apply/match this pattern in /each line/ of this multi-line string.' ?

Answer 1

It looks like you're trying to use square brackets for grouping, which is wrong. Square brackets define a character class, within which most characters lose any special meanings they normally have, but class shorthands like \s and \S still work. So you've created a character class that matches parentheses, plus signs, any whitespace character, and anything that's not a whitespace character. In other words, it matches everything.
The regex you need is actually very simple:  Pattern commentPattern = Pattern.compile("##.*");I should also mention that the '$' anchor normally doesn't match line ends; you have to compile the Pattern with the MULTILINE flag set to make it do that. It wouldn't make a difference with your regex, though, since the [\s\S]+ will always gobble up the rest of the input anyway.