CODEWITHSUNDEEP
rubyregexgitlabre2
Moon
Moon

Reputation: 35305

Regex to filter out word ending with a “string” suffix - without lookahead

I am trying to come up with a Ruby Regex that will match the following string:

MAINT: Refactor something
STRY-1: Add something
STRY-2: Update something

But should not match the following:

MAINT: Refactored something
STRY-1: Added something
STRY-2: Updated something

MAINT: Refactoring something
STRY-3: Adding something
STRY-4: Updating something

Basically, the first word after : should not end with either ed or ing

I have been using the following regex for GitLab commit message for a while now.

^(MAINT|(STRY|PRB)-\d+):\s(?:(?!(?:ed|ing)\b)[A-Za-z])+\s([a-zA-Z0-9._\-"].*)

However, recently they seem to have switched to using google/re2 which does not support lookahead.

Would it be possible to rewrite this regex in way so that lookahead is not used?

Upvotes: 0

Views: 1067

Answers (3)

Cary Swoveland
Cary Swoveland

Reputation: 110725

str =<<_
MAINT: Refactor something
STRY-1: Added something
MAINT: Refactoring something
Add something
STRY-3: Adding something
STRY-1: Add something
MAINT: Refactored something
Refactor something
STRY-4: Updating something
STRY-9:   Update something
STRY-2: Updated something
_

r = /
    ^                      # Match beginning of line
    (?:                    # Begin non-capture group
      MAINT\:[ ]+Refactor  # Match string
      |                    # or
      STRY-\d+\:[ ]+       # match string
      (?:Add|Update)       # match 'Add' or 'Update'
    )                      # end non-capture group
    [ ]+something          # match one or more spaces followed by 'something'
    $                      # match end of line
    /x                     # free-spacing regex definition modes

str.scan(r)
  #=> ["MAINT: Refactor something\n",
  #    "STRY-1: Add something\n",
  #    "STRY-9:   Update something\n"]

To match a space in the regular expression I've use a character class containing a space ([ ]). That's needed because free-spacing mode removes spaces that are not in character classes. Written in the convention way, the regular expression is as follows.

/^(?:MAINT\: +Refactor|STRY-\d+\: +(?:Add|Update)) +something$/

Upvotes: 1

revo
revo

Reputation: 48741

You are dealing with a regex which has to be aware about three endings:

  1. ed\b
  2. ing\b
  3. ied\b

You have to consider existence of each single spot. For instance, e[^d]\b and [^e]d\b. Writing all of them you will come with this regex: 

^(MAINT|(STRY|PRB)-\d+):\s*(?i:\w*(e[a-ce-z]|[a-df-z]d|i(n[a-fh-z]|[a-mo-z]g|e[a-ce-z]|[a-df-z]d)|[a-hj-z]ng|[a-hj-z][a-df-mo-z][a-cefh-z])|\w)\s([a-zA-Z0-9._\-"].*)

Live demo

Upvotes: 1

steenslag
steenslag

Reputation: 80085

Without regular expression:

str = "MAINT: Refactor something
STRY-1: Add something
STRY-2: Update something"

p str.lines.none?{|line| line.split[1].end_with?("ed", "ing")}
# => true

Upvotes: -1

Related Questions