in reply to Re^4: Help with Parse::RecDescent grammar
in thread Help with Parse::RecDescent grammar

Here is the output I get:

Oops, portList should be 

   portList : "'" "(" record(s?) ")"
              { $item[3] }

[download]

I made a minor change (undid my mergine of ' and ( into a single token) right before posting without testing.

By the way, "'(" means no whitespace (as defined by <skip>, \s* by default) is allwed between the two characters, while "'" "(" means whitespace IS allowed.

could this have something to do with the #f?

Is that a commment? I treated it as a comment, so I used <skip> to handle it.

If it's not a comment, what is it? An unquoted string? Does it support any kind of escapes?

Replies are listed 'Best First'.
Re^6: Help with Parse::RecDescent grammar
by dramguy (Novice) on Dec 13, 2006 at 14:25 UTC
    The  #f is a boolean value used to tell a EDA program how to update the corresponding data. The #f, tells the system to only update new data, whereas #t tells the system to overwrite all previous data. This always comes after the last ")" and is never quoted. Im not sure what you would call this?

    I changed the code to return $item[3] and I now get a nice looking structure. One thing I notice is when I try to parse multiple entries, I get the bad netlist error again. For example:

    Consider my data is the following: 

    my $text = <<'_EOT_';
dbSetCellPortTypes "/opt/mylib/s956M" "blahblah" '(
  ("gnd!" "Inout" "Ground" )
  ("vint!" "Inout" "Power" )
) #f
dbSetCellPortTypes "/opt/mylib/s956M" "newCell" '(
  ("gnd!" "Inout" "Ground" )
  ("vint!" "Inout" "Power" )
) #f
_EOT_

    [download]

    I've tried changing the parse rule to the following as well: 

     parse    : <skip:'(?:\s*(?:#[^\n]*)?)'> portDef(s?) /\Z/
              { $item[2] }

    [download]

    What is the /\Z/ doing actually?

    Thanks,

    Frank

[reply]
[d/l]
[select]

      The #f is a boolean value

      ah! 

      my $grammar = <<'_EOGRAMMAR_';
   {
      # These apply to code in those block and all actions.
      use strict;
      use warnings;

      my %escapes = (
         n => "\n",
      );

      sub dequote_double {
         for (my $s = @_ ? $_[0] : $_) {
            s/^"//g;
            s/"$//g;
            s/\\(.)/$escapes{$1} || $1/eg;
            return $_;
         }
      }
   }

   parse    : portDef(s?) /\Z/
              { $item[1] }

   portDef  : "dbSetCellPortTypes" QSTRING QSTRING portList BOOL
              { [ @item[1..5] ] }

   portList : "'" "(" record(s?) ")"
              { $item[3] }

   record   : "(" QSTRING(s?) ")"
              { $item[2] }

   QSTRING  : /"(?:[^"\\]|\\.)*"/
              { dequote_double($item[1]) }

   BOOL     : /#[tf]\b/

_EOGRAMMAR_

      [download]

      One thing I notice is when I try to parse multiple entries, I get the bad netlist error again.

      Copy and paste error. There was a missing *.

      What is the /\Z/ doing actually?

      /\Z/ means makes sure we're at end of file. Without it, extra garbage at the end will just get ignored. It's worst that it sounds, because anything the parser doesn't understand is garbage, includes malformed dbSetCellPortTypes records. If there were a syntax error halfway through the file, you'd silently ignore everything after the syntax error if the /\Z/ is missing.
[reply]
[d/l]
[select]
        Ahh, Ok, I think Im starting to understand how this module works.

        One more question though. Im trying to skip all lines which begin with a ";" as this is a comment. I used the following code to no avail: 

        <skip:';'>

        [download]

        Am I missing something here?

        Thanks

        F

[reply]
[d/l]

In Section Seekers of Perl Wisdom