The Perl Toolchain Summit needs more sponsors. If your company depends on Perl, please support this very important event.
#!/usr/bin/perl
#
# eyapp -- Front end to the Parse::Eyapp module
#
#

=head1 NAME

eyapp - A Perl front-end to the Parse::Eyapp module


=head1 SYNOPSYS

    eyapp [options] grammar[.eyp]
    eyapp -V
    eyapp -h

        grammar     The grammar file. If no suffix is given, and the file
                    does not exists, .eyp is added

=head1 DESCRIPTION

The eyapp compiler is a front-end to the L<Parse::Eyapp> module, which lets you compile
Parse::Eyapp grammar input files into Perl LALR(1) Object Oriented parser modules.


=head1 OPTIONS IN DETAIL

=over 4

=item I<-v>

Creates a file F<grammar>.output describing your parser. It will
show you a summary of conflicts, rules, the DFA (Deterministic
Finite Automaton) states and overall usage of the parser.

Implies option C<-N>. To produce a more detailed description of the states,
the LALR tables aren't compacted.
Use the combination C<-vN> to produce an  C<.output> file 
corresponding to the compacted tables.


=item I<-s>

Create a standalone module in which the parsing driver is included.
The modules including the LALR driver (L<Parse::Eyapp::Driver>),
those for AST manipulations (L<Parse::Eyapp::Node> and
L<Parse::Eyapp::YATW>)) and L<Parse::Eyapp::Base> 
are included - almost verbatim - inside the generated module. 

Note that if you have more than one parser module called from a program, 
to have it standalone, you need this option only for one of your grammars;

=item I<-n>

Disable source file line numbering embedded in your parser module.
I don't know why one should need it, but it's there.

=item I<-m module>

Gives your parser module the package name (or name space or module name or
class name or whatever-you-call-it) of F<module>.  It defaults to F<grammar>

=item I<-o outfile>

The compiled output file will be named F<outfile> for your parser module.
It defaults to F<grammar>.pm or, if you specified the option
I<-m A::Module::Name> (see below), to F<Name.pm>.

=item I<-c grammar>[.eyp]

Produces as output (STDOUT) the grammar without the actions. Only the syntactic
parts are displayed. Comments will be also stripped 
if the C<-v> option is added.


=item I<-t filename>

The I<-t filename> option allows you to specify a file which should be 
used as template for generating the parser output.  The default is to 
use the internal template defined in F<Parse::Eyapp::Output.pm>.
For how to write your own template and which substitutions are available,
have a look to the module F<Parse::Eyapp::Output.pm> : it should be obvious. 

=item I<-b shebang>

If you work on systems that understand so called I<shebangs>, and your
generated parser is directly an executable script, you can specify one
with the I<-b> option, ie:

    eyapp -b '/usr/local/bin/perl -w' -o myscript.pl myscript.yp

This will output a file called F<myscript.pl> whose very first line is:

    #!/usr/local/bin/perl -w

The argument is mandatory, but if you specify an empty string, the value
of I<$Config{perlpath}> will be used instead.

=item I<-B prompt>   

Adds a modulino call '__PACKAGE->main(<prompt>) unless caller();' 
as the very last line of the output file. The argument is mandatory.


=item  I<-C grammar.eyp>         

An abbreviation for the combined use of I<-b ''> and  I<-B ''>

=item I<-T grammar.eyp>         

Equivalent to C<%tree>.


=item I<-N grammar.eyp>         

Equivalent to the directive C<%nocompact>. Do not compact LALR 
action tables. 


=item I<-l>         

Do not provide a default lexical analyzer. By default C<eyapp>
builds a lexical analyzer from your C<%token = /regexp/> definitions

=item I<grammar>

The input grammar file. If no suffix is given, and the file does not exists,
an attempt to open the file with a suffix of  F<.eyp> is tried before exiting.

=item I<-V>

Display current version of Parse::Eyapp and gracefully exits.

=item I<-h>

Display the usage screen.

=back

=head1 EXAMPLE

The following C<eyapp> program translates an infix expression
like C<2+3*4> to postfix: C<2 3 4 * +>

    %token NUM = /([0-9]+(?:\.[0-9]+)?)/
    %token VAR = /([A-Za-z][A-Za-z0-9_]*)/

    %right  '='
    %left   '-' '+'
    %left   '*' '/'
    %left   NEG

    %defaultaction { "$left $right $op"; }

    %%
    line: $exp  { print "$exp\n" }
    ;

    exp:        $NUM  { $NUM }            
            |   $VAR  { $VAR }            
            |   VAR.left '='.op exp.right         
            |   exp.left '+'.op exp.right         
            |   exp.left '-'.op exp.right        
            |   exp.left '*'.op exp.right       
            |   exp.left '/'.op exp.right      
            |   '-' $exp %prec NEG { "$exp NEG" }
            |   '(' $exp ')' { $exp }      
    ;

    %%

Notice that there is no need to write lexer and error report subroutines.
First, we compile the grammar:

    pl@nereida:~/LEyapp/examples/eyappintro$ eyapp -o postfix.pl -C Postfix.eyp 

If we use the C<-C> option and no C<main()> was written one default C<main> sub is provided.
We can now execute the resulting program:

    pl@nereida:~/LEyapp/examples/eyappintro$ ./postfix.pl -c 'a = 2*3 +b'
    a 2 3 * b + =

When a non conformant input is given, it produces an accurate error message:

    pl@nereida:~/LEyapp/examples/eyappintro$ ./postfix.pl -c 'a = 2**3 +b'

    Syntax error near '*'. 
    Expected one of these terminals: '-' 'NUM' 'VAR' '('
    There were 1 errors during parsing


=head1 AUTHOR

Casiano Rodriguez-Leon 

=head1 COPYRIGHT

(c) Copyright 2006 Casiano Rodriguez-Leon

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.

=head1 SEE ALSO

=over

=item * L<Parse::Eyapp>,

=item *
perldoc L<vgg>,

=item * The tutorial I<Parsing Strings and Trees with> C<Parse::Eyapp>
(An Introduction to Compiler Construction in seven pages)> in

=item * The pdf file in L<http://nereida.deioc.ull.es/~pl/perlexamples/Eyapp.pdf> 

=item * L<http://nereida.deioc.ull.es/~pl/perlexamples/section_eyappts.html> (Spanish),

=item * L<eyapp>,

=item * L<treereg>,

=item * L<Parse::yapp>,

=item * yacc(1),

=item * bison(1),

=item * the classic book "Compilers: Principles, Techniques, and Tools" by Alfred V. Aho, Ravi Sethi and

=item * Jeffrey D. Ullman (Addison-Wesley 1986)

=item * L<Parse::RecDescent>.

=back



=cut

require 5.004;

use File::Basename;
use Getopt::Std;
use Config;
use Parse::Eyapp::Base qw(compute_lines);
use Parse::Eyapp;

use strict;

our ( $opt_n, $opt_m, $opt_V, $opt_v, $opt_o, $opt_h, $opt_s, $opt_t, $opt_b, $opt_c, $opt_B, $opt_C, );
our $opt_S; # Specify start symbol
our $opt_l = 0;
our $opt_T; # build AST
our $opt_N; # Do not compact action tables. No DEFAULT field for "STATES"
our $opt_P; # Accept if prefix conforms to the language, even if not end of input
our $opt_w; #  Produce a .dot graph describing the LR non-compacted automaton
our $opt_W; #  Produce a .dot graph describing the LR non-compacted automaton

sub Usage {
    my($prog)=(fileparse($0,'\..*'))[0];
    die <<EOF;

Usage:  $prog [options] grammar[.eyp]
  or    $prog -V
  or    $prog -h

    -m module   Give your parser module the name <module>
                default is <grammar>
    -v          Create a file <grammar>.output describing your parser
                The LALR tables aren't compacted
    -vN         Create a file <grammar>.output describing your parser
                LALR tables are compacted.
    -w          Create a file <grammar>.dot graph describing your parser.
                You can then generate a .png file using dot:
                            dot -Tpng  grammar.dot -o grammar.png
    -W          Create a file <grammar>.dot graph describing your parser
                Nodes are labelled with core LR-items
    -s          Create a standalone module in which the driver is included
    -S symbol   Use 'symbol' as start symbol of the grammar
    -n          Disable source file line numbering embedded in your parser
    -o outfile  Create the file <outfile> for your parser module
                Default is <grammar>.pm or, if -m A::Module::Name is
                specified, Name.pm
    -t filename Uses the file <filename> as a template for creating the parser
                module file.  Default is to use internal template defined
                in Parse::Eyapp::Output
    -b shebang  Adds '#!<shebang>' as the very first line of the output file
    -B prompt   Adds a modulino call '__PACKAGE->main(<prompt>) unless caller();' 
                as the very last line of the output file
    -C          An abbreviation for the combined use of -b ''  -B ''
    -T          Equivalent to %tree
    -N          Equivalent to %nocompact. Do not compact action tables. 
    -l          Do not generate the lexical analyzer. One must be explictly provided
    -P          Accept if prefix conforms to the language, even if is not legal up 
                to the end 
    -c grammar  Displays the "clean" grammar without actions     
    -vc grammar Displays the "clean" grammar without actions. Strip comments also
    -V          Display current version of Parse::Eyapp and gracefully exits
    -h          Display this help screen

    grammar     The grammar file. If no suffix is given, and the file
                does not exists, .eyp is added

EOF
}

my($nbargs)=@ARGV;

    getopts('CTVNhlvsnwWPc:b:B:m:t:o:S:')
or  Usage;

   (  ($opt_V and $nbargs > 1)
    or $opt_h)
and Usage;

    $opt_V
and do {

    @ARGV == 0 or  Usage;

    print "This is Parse::Eyapp version $Parse::Eyapp::Driver::VERSION.\n";
    exit(0);

};

($opt_w or $opt_W) and do { 
  $opt_v = 1;
};

    $opt_c and do {
     my $file;

     local $/ = undef;
     open($file, $opt_c) or die "Cannot open grammar file $opt_c: $!\n";
     $opt_c = <$file>;
     close($file);

     require Parse::Eyapp::Cleaner;
     my @args = $opt_v ? (skipcomments => 1) : ();
     print Parse::Eyapp::Cleaner::ppcontroller(\$opt_c, @args);
     exit(0);
  };



# -t <filename> ($opt_t) option allows a file to be specified which 
# contains a 'template' to be used when generating the parser; 
# if defined, we open and read the file.   

    $opt_t
and do {
    local $/ = undef;
    local *TFILE;
    open(TFILE, $opt_t)
    or die "Cannot open template file $opt_t: $!\n";
    $opt_t = <TFILE>;
    close(TFILE);
};

    @ARGV == 1
or  Usage;

my($filename)=$ARGV[0];
my($base,$path,$sfx)=fileparse($filename,'\..*');

    -r "$filename"
or  do {
        ($sfx eq '.yp' or $sfx eq '.eyp')
    or (-r "$filename.yp" and $filename.='.yp')
    or (-r "$filename.eyp" and $filename.='.eyp');

        -r "$filename"
    or  die "Cannot open $filename for reading.\n";
};

my @args = (inputfile => $filename, );

push @args, (tree => 1) if $opt_T;
push @args, (nocompact => 1) if (($opt_N or $opt_v) and not($opt_N and $opt_v));
push @args, (prefix => 1) if $opt_P;
push @args, (start => $opt_S) if $opt_S;

my ($parser) = Parse::Eyapp->new(@args);

my($warnings)=$parser->Warnings();

$warnings !~ m{^\s*$} and print STDERR "$warnings\n";

$parser->outputtables($path, $base) if $opt_v;

if ($opt_w or $opt_W) { 
  $parser->outputDot($path, $base, 0) if $opt_w;
  $parser->outputDot($path, $base, 1) if $opt_W;

  my ($pngfile)= "$path$base.png";
  my ($dotfile)= "$path$base.dot";
  system qq{dot -Tpng $dotfile -o $pngfile};
}

my($outfile)="$path$base.pm";
my($package)="$base";

# Use the name of the start symbol for the package, if specified
$opt_m = $opt_S unless $opt_m;

    $opt_m
and do {
    $package=$opt_m;
    $package=~/^(?:
                  (?:[^:]|:(?!:))*::)*       # The ID up to the last ::
                  (.*)                       # No colons here
               $/x;
    $outfile="$1.pm";
};

    $opt_o
and $outfile=$opt_o;

$opt_s = $opt_s ? 1 : 0;

$opt_n = $opt_n ? 0 : 1;

    open(OUT,">$outfile")
or  die "Cannot open $outfile for writing.\n";

    defined($opt_C) and do {
      $opt_b = '' unless defined $opt_b;
      $opt_B = '' unless defined $opt_B;
    };

    defined($opt_b)
and do {
        $opt_b
    or  $opt_b = $Config{perlpath};
    print OUT "#!$opt_b\n";
};

my $text = $parser->Output(
                          classname  => $package,
                          standalone => $opt_s,
                          linenumbers => $opt_n,
                          template    => $opt_t,
                          modulino    => $opt_B,
                          lexerisdefined => $opt_l,
                          #prefixname => 'R::',
                         );
compute_lines(\$text, $outfile, $Parse::Eyapp::Output::pattern) if $opt_n;

print OUT $text;

close(OUT);

chmod(0755, $outfile) if $opt_b;