treereg(1) - Linux man page

Name

treereg - Compiler for Tree Regular Expressions

Synopsis

treereg [-m packagename] [[no]syntax] [[no]numbers] [-severity 0|1|2|3] \
        [-p treeprefix] [-o outputfile] [-lib /path/to/library/] -i filename[.trg]
treereg [-m packagename] [[no]syntax] [[no]numbers] [-severity 0|1|2|3] \
        [-p treeprefix] [-lib /path/to/library/] [-o outputfile] filename[.trg]
treereg -v
treereg -h

Options

Options can be used both with one dash and double dash. It is not necessary to write the full name of the option. A disambiguation prefix suffices.

• "-i[n] filename"

Input file. Extension ".trg" is assumed if no extension is provided.

• "-o[ut] filename"

Output file. By default is the name of the input file (concatenated with .pm)

• "-m[od] packagename"

Name of the package containing the generated subroutines. By default is the longest prefix of the input file name that conforms to the classic definition of integer "[a-z_A-Z]\w*".

• "-l[ib] /path/to/library/"

Specifies that "/path/to/library/" will be included in @INC. Useful when the "syntax" option is on. Can be inserted as many times as necessary.

• "-p[refix] treeprefix"

Tree nodes automatically generated using "Parse::Eyapp" are objects blessed into the name of the production. To avoid crashes the programmer may prefix the class names with a given prefix when calling the parser; for example:

$self->YYParse( yylex => \&_Lexer, yyerror => \&_Error, yyprefix => __PACKAGE__."::")

The "-prefix treeprefix" option simplifies the process of writing the tree grammar so that instead of writing with the full names

CLASS::TIMES(CLASS::NUM, $x) and { $NUM->{VAL} == 0) => { $NUM }

it can be written:

TIMES(NUM, $x) and { $NUM->{VAL} == 0) => { $NUM }

• "-n[umbers]"

Produces "#line" directives.

• "-non[umbers]"

Disable source file line numbering embedded in your parser

• "-sy[ntax]"

Checks that Perl code is syntactically correct.

• "-nosy[ntax]"

Does not check the syntax of Perl code

• "-se[verity] number"

- 0 = Don't check arity (default). Matching does not check the arity. The actual node being visited may have more children.

- 1 = Check arity. Matching requires the equality of the number of children and the actual node and the pattern.

- 2 = Check arity and give a warning

- 3 = Check arity, give a warning and exit

• "-v[ersion]"

Gives the version

• "-u[sage]"

Prints the usage info

• "-h[elp]"

Print this help

Description

"Treereg" translates a tree grammar specification file (default extension ".trg" describing a set of tree patterns and the actions to modify them using tree-terms like:

TIMES(NUM, $x) and { $NUM->{VAL} == 0) => { $NUM }

which says that wherever an abstract syntax tree representing the product of a numeric expression with value 0 times any other kind of expression, the "TIMES" tree can be substituted by its left child.

The compiler produces a Perl module containing the subroutines implementing those sets of pattern-actions.

Example

Consider the following "eyapp" grammar (see the "Parse::Eyapp" documentation to know more about "Parse::Eyapp" grammars):

----------------------------------------------------------
nereida:~/LEyapp/examples> cat Rule6.yp
%{
use Data::Dumper;
%}
%right  '='
%left   '-' '+'
%left   '*' '/'
%left   NEG
%tree

%%
line: exp  { $_[1] }
;

exp:      %name NUM
            NUM
        | %name VAR
          VAR
        | %name ASSIGN
          VAR '=' exp
        | %name PLUS
          exp '+' exp
        | %name MINUS
          exp '-' exp
        | %name TIMES
          exp '*' exp
        | %name DIV
          exp '/' exp
        | %name UMINUS
          '-' exp %prec NEG
        |   '(' exp ')'  { $_[2] } /* Let us simplify a bit the tree */
;

%%

sub _Error {
    die  "Syntax error.\n";
}

sub _Lexer {
    my($parser)=shift;

        $parser->YYData->{INPUT}
    or  $parser->YYData->{INPUT} = <STDIN>
    or  return('',undef);

    $parser->YYData->{INPUT}=~s/^\s+//;

    for ($parser->YYData->{INPUT}) {
        s/^([0-9]+(?:\.[0-9]+)?)// and return('NUM',$1);
        s/^([A-Za-z][A-Za-z0-9_]*)// and return('VAR',$1);
        s/^(.)//s and return($1,$1);
    }
}

sub Run {
    my($self)=shift;
    $self->YYParse( yylex => \&_Lexer, yyerror => \&_Error );
}
----------------------------------------------------------

Compile it using "eyapp":

----------------------------------------------------------
nereida:~/LEyapp/examples> eyapp Rule6.yp
nereida:~/LEyapp/examples> ls -ltr | tail -1
-rw-rw----  1 pl users  4976 2006-09-15 19:56 Rule6.pm
----------------------------------------------------------

Now consider this tree grammar:

----------------------------------------------------------
nereida:~/LEyapp/examples> cat Transform2.trg
%{
my %Op = (PLUS=>'+', MINUS => '-', TIMES=>'*', DIV => '/');
%}

fold: 'TIMES|PLUS|DIV|MINUS':bin(NUM($n), NUM($m))
  => {
    my $op = $Op{ref($bin)};
    $n->{attr} = eval  "$n->{attr} $op $m->{attr}";
    $_[0] = $NUM[0];
  }
zero_times_whatever: TIMES(NUM($x), .) and { $x->{attr} == 0 } => { $_[0] = $NUM }
whatever_times_zero: TIMES(., NUM($x)) and { $x->{attr} == 0 } => { $_[0] = $NUM }

/* rules related with times */
times_zero = zero_times_whatever whatever_times_zero;
----------------------------------------------------------

Compile it with "treereg":

----------------------------------------------------------
nereida:~/LEyapp/examples> treereg Transform2.trg
nereida:~/LEyapp/examples> ls -ltr | tail -1
-rw-rw----  1 pl users  1948 2006-09-15 19:57 Transform2.pm
----------------------------------------------------------

The following program makes use of both modules "Rule6.pm" and "Transform2.pm":

----------------------------------------------------------
nereida:~/LEyapp/examples> cat foldand0rule6_3.pl
#!/usr/bin/perl -w
use strict;
use Rule6;
use Parse::Eyapp::YATW;
use Data::Dumper;
use Transform2;

$Data::Dumper::Indent = 1;
my $parser = new Rule6();
my $t = $parser->Run;
print "\n***** Before ******\n";
print Dumper($t);
$t->s(@Transform2::all);
print "\n***** After ******\n";
print Dumper($t);
----------------------------------------------------------

When the program runs with input "b*(2-2)" produces the following output:

----------------------------------------------------------
nereida:~/LEyapp/examples> foldand0rule6_3.pl
b*(2-2)

***** Before ******
$VAR1 = bless( {
  'children' => [
    bless( {
      'children' => [
        bless( { 'children' => [], 'attr' => 'b', 'token' => 'VAR' }, 'TERMINAL' )
      ]
    }, 'VAR' ),
    bless( {
      'children' => [
        bless( { 'children' => [
            bless( { 'children' => [], 'attr' => '2', 'token' => 'NUM' }, 'TERMINAL' )
          ]
        }, 'NUM' ),
        bless( {
          'children' => [
            bless( { 'children' => [], 'attr' => '2', 'token' => 'NUM' }, 'TERMINAL' )
          ]
        }, 'NUM' )
      ]
    }, 'MINUS' )
  ]
}, 'TIMES' );

***** After ******
$VAR1 = bless( {
  'children' => [
    bless( { 'children' => [], 'attr' => 0, 'token' => 'NUM' }, 'TERMINAL' )
  ]
}, 'NUM' );
----------------------------------------------------------

See also the section "Compiling: More Options" in Parse::Eyapp for a more contrived example.

Author

Casiano Rodriguez-Leon

License And Copyright

This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself, either Perl version 5.8.8 or, at your option, any later version of Perl 5 you may have available.