Verilog::Parser(3) Parse Verilog language files

SYNOPSIS


use Verilog::Parser;
my $parser = new Verilog::Parser;
$string = $parser->unreadback ();
$line = $parser->lineno ();
$parser->parse ($text)
$parser->parse_file ($filename)

DESCRIPTION

Verilog::Parser will tokenize a Verilog file when the parse() method is called and invoke various callback methods. This is useful for extracting information and editing files while retaining all context. For netlist like extractions, see Verilog::Netlist.

See the ``Which Package'' section of Verilog::Language if you are unsure which parsing package to use for a new application.

Note the parser allows some constructs that are syntax errors according to the specification (for example ``foo.bar(1)++''.) This is done when the parser can't easily detect these cases. It's up to the consumer of the parser to filter out such errors if it cares.

METHODS

$parser = Verilog::Parser->new (args...)
Create a new Parser.

Adding ``symbol_table => []'' will use the specified symbol table for this parse, and modify the array reference to include those symbols detected by this parse. As the SystemVerilog language requires packages and typedefs to exist before they are referenced, you must pass the same symbol_table to subsequent parses that are for the same compilation scope. The internals of this symbol_table should be considered opaque, as it will change between package versions, and must not be modified by user code.

Adding ``use_cb_{callback-name} => 0'' will disable the specified callback. By default, all callbacks will be called; disabling callbacks can greatly speed up the parser as a large percentage of time is spent calling between C and Perl to invoke the callbacks. When using this feature, use_unreadback=>0 should be used too, as since whole tokens are skipped, skipping whitespace shouldn't matter either.

Adding ``use_protected => 0'' will disable callbacks on `protected regions, which may improve performance.

Adding ``use_std => 1'' will add parsing of the SystemVerilog built-in std:: package, or ``use_std => 0'' will disable it. If unspecified it is silently included (no callbacks will be involed) when suspected to be necessary.

Adding ``use_unreadback => 0'' will disable later use of the unreadback method, which may improve performance.

Adding ``use_vars => 0'' will disable contassign, defparam, pin, var and port callbacks to Verilog::SigParser. This can greatly speed parsing when variable and interconnect information is not required.

$parser->callback_names ()
Return an array of callback function names. This may be used to automatically create callbacks for all functions, or to test for different callback functionality between versions of Verilog-Perl.
$parser->eof ()
Indicate the end of the input stream. All incomplete tokens will be parsed and all remaining callbacks completed.
$parser->filename ($set)
Return (if $set is undefined) or set current filename.
$parser->lineno ($set)
Return (if $set is undefined) or set current line number.
$parser->parse ($string)
Parse the $string as verilog text. Can be called multiple times. Note not all callbacks may be invoked until the eof method is called.
$parser->parse_file ($filename);
This method can be called to parse text from a file. The argument can be a filename or an already opened file handle. The return value from parse_file() is a reference to the parser object.
$parser->parse_preproc_file ($preproc);
This method can be called to parse preprocessed text from a predeclared Verilog::Preproc object.
$parser->unreadback ($string)
Return any input string from the file that has not been sent to the callback. This will include whitespace and tokens which did not have a callback. (For example comments, if there is no comment callback.) This is useful for recording the entire contents of the input, for preprocessors, pretty-printers, and such.

With the optional argument, set the text to be returned with the next unreadback call. See also unreadbackCat, which is much faster.

To use this option, ``use_unreadback => 1'' must have been passed to the constructor.

$parser->unreadbackCat ($text)
Add text to be returned with the next unreadback call. This is much faster than using ``$parser->unreadback($parser->unreadback . $text)''.

CALLBACKS

In order to make the parser do anything interesting, you must make a subclass where you override one or more of the following callback methods as appropriate.
$self->attribute ( $token )
This method is called when any text in (* *) are recognized. The first argument, $token, is the contents of the attribute including the delimiters.
$self->comment ( $token )
This method is called when any text in // or /**/ comments are recognized. The first argument, $token, is the contents of the comment including the comment delimiters.
$self->endparse ( $token )
This method is called when the file has been completely parsed, at the End-Of-File of the parsed file. It is useful for writing clean up routines.
$self->keyword ( $token )
This method is called when any Verilog keyword is recognized. The first argument, $token, is the keyword.
$self->number ( $token )
This method is called when any number is recognized. The first argument, $token, is the number. The Verilog::Language::number_value function may be useful for converting a Verilog value to a Perl integer.
$self->operator ( $token )
This method is called when any symbolic operator (+, -, etc) is recognized. The first argument, $token, is the operator.
$self->preproc ( $token )
This method is called when any Verilog preprocessor `command is recognized. Most of these are handled by the preprocessor, however any unrecognized `defines are passed through. For backward compatibility, if not defined this function will call the symbol callback.
$self->string ( $token )
This method is called when any text in double quotes are recognized, or on the text of protected regions. The first argument, $token, is the contents of the string including the quotes.
$self->symbol ( $token )
This method is called when any Verilog symbol is recognized. A symbol is considered a non-keyword bare-word. The first argument, $token, is the symbol.
$self->sysfunc ( $token )
This method is called when any Verilog $syscall is recognized. The first argument, $token, is the symbol. For backward compatibility, if not defined this function will call the symbol callback.

EXAMPLE

Here's a simple example which will print every symbol in a verilog file.

  package MyParser;
  use Verilog::Parser;
  @ISA = qw(Verilog::Parser);
  # parse, parse_file, etc are inherited from Verilog::Parser
  sub new {
      my $class = shift;
      #print "Class $class\n";
      my $self = $class->SUPER::new();
      bless $self, $class;
      return $self;
  }
  sub symbol {
      my $self = shift;
      my $token = shift;
      $self->{symbols}{$token}++;
  }
  sub report {
      my $self = shift;
      foreach my $sym (sort keys %{$self->{symbols}}) {
         printf "Symbol %-30s occurs %4d times\n",
         $sym, $self->{symbols}{$sym};
      }
  }
  package main;
  my $parser = MyParser->new();
  $parser->parse_file (shift);
  $parser->report();

BUGS

This is being distributed as a baseline for future contributions. Don't expect a lot, the Parser is still naive, and there are many awkward cases that aren't covered.

The parser currently assumes the string it is passed ends on a newline boundary. It should be changed to allow arbitrary chunks.

Cell instantiations without any arguments are not supported, an empty set of parenthesis are required. (Use ``cell cell();'', not ``cell cell;''.)

DISTRIBUTION

Verilog-Perl is part of the <http://www.veripool.org/> free Verilog EDA software tool suite. The latest version is available from CPAN and from <http://www.veripool.org/verilog-perl>.

Copyright 2000-2016 by Wilson Snyder. This package is free software; you can redistribute it and/or modify it under the terms of either the GNU Lesser General Public License Version 3 or the Perl Artistic License Version 2.0.

AUTHORS

Wilson Snyder <[email protected]>