Mirror of forum.nim-lang.org

5804 :: stdlib pegs: Accessing optional subexpressions

• sschwarzer (orginal) [2020-01-12T19:17:21+01:00] view original

  Assume I have the following PEG definition:

  import pegs
  
  let examplePeg = peg"""
    A <- {L}? {D}?
    L <- \a+
    D <- \d+
  """
  Run

  If I test strings against this PEG, I get:

  if "abc" =~ examplePeg:
    # `["abc", "", "", "", "", "", "", "", "", "", "", "", "", "", "", "", "", "", "", ""]`
    echo matches
  
  if "123" =~ examplePeg:
    # `["123", "", "", "", "", "", "", "", "", "", "", "", "", "", "", "", "", "", "", ""]
    echo matches
  Run

  Now, since the subexpressions L and D are optional, it seems I need to look at the matched strings and check them again despite the PEG already defines whether I expect letters or digits for these subexpressions.

  The only API I found that gives more control is eventParser, which on the other hand seems far too complicated for a relatively simple case as this.

  Is there a simpler API (or a combination of them) in the pegs module, which I may have missed?

---

• sky_khan (orginal) [2020-01-13T05:22:58+01:00] view original

  Afaik, stdlib peg is designed for nim compiler itself and somewhat limited. So, you may consider using npeg :

  import npeg
  
  let parser = peg("example"):
    example <- >?word * >?number
    word <- +Alpha
    number <- +Digit
  
  echo parser.match("abc").captures # @["abc", ""]
  echo parser.match("123").captures # @["", "123"]
  Run

---

• sschwarzer (orginal) [2020-01-13T22:10:24+01:00] view original

  Thanks!

  I had heard of npegs, but so far chose pegs because it's in the standard library and understands standard PEG syntax. That said, npeg's features are very impressive. :-)

  Afaik, stdlib peg is designed for nim compiler itself and somewhat limited.

  So I guess there's not much motivation to enhance this module? Probably it would be easier to implement a standard PEG parser with npegs as part of npegs than enhance the pegs module? ;-)

---