CODEWITHSUNDEEP
regexperlwhitespaceremoving-whitespace
JonMorehouse
JonMorehouse

Reputation: 1383

Perl MultiLine Regex Statement

I'm having some trouble getting a multiline regex statement to work for me.

Basically, I'm trying to remove empty lines following a ) ->. Matching the multiline section has been a bit tricky. Here's what I have so far: 

perl -00 -p -i -e 's/\) ->(?=[^\n]*\n\n)$/\) ->\n/m' $filename

Here's my input/output:

Input:
setUp = 
  config: (cb) ->

    randomFunction (cb)->

       cb?()

  nestedObject: 
    key: (cb) ->

      cb?()


Output:
setUp = 
  config: (cb) ->
    cb?()

  nestedObject: 
    key: (cb) ->
      cb?()

Upvotes: 2

Views: 192

Answers (3)

Casimir et Hippolyte
Casimir et Hippolyte

Reputation: 89557

You can use this replacement:

s/\) ->\R\K\R+//g

\R is a shortcut for an atomic group that contains several common types of newlines
\K removes all on the left from match result

Upvotes: 2

Miller
Miller

Reputation: 35198

Just do line by line processing with a flip-flop range, removing all blank lines as the end condition:

perl -i -pe '/\) ->\s*$/...!s/^\s*$//' file.txt

Perhaps a little easier to read:

perl -i -pe 'm{\) ->\s*$}...!s/^\s*$//' file.txt

Upvotes: 2

mob
mob

Reputation: 118605

Move the newline characters out of the look-ahead. Try

    s/\) ->(?=[^\n]*)\n\n/\) ->\n/mg;

Characters in the look-ahead are not replaced in a substitution.

(Actually, I don't see why you even need a look-ahead.

    s/\) ->.*\n\n/\) ->\n/mg;

also does the job, and any non-zero length sequence that matched the look-ahead would also make the whole pattern match fail.)

You also may want to use the /g flag, since you want to do this substitution more than once in the document.

Upvotes: 2

Related Questions