SGML::Parser::OpenSP::Tools - Tools to process OpenSP output


Routines to post-process OpenSP event data.



specified_attribute returns a true value if the attribute is of type cdata or tokenized and has its Defaulted property set to specified. For example

  sub start_element
    my $self = shift;
    my $elem = shift;
    my @spec = grep specified_attribute($_),
                    values %{$elem->{Attributes}};

    # @spec contains all explicitly specified attributes

defaulted_attribute returns a true value if the attribute is of type cdata or tokenized and has its Defaulted property set to something but specified. For all attributes, the following always holds true,

  !defined(attribute_value($_)) or
  defaulted_attribute($_) or

since only defaulted and specified attributes can have a value.


Returns true if the value can have a value, i.e., it is either specified or defaulted.


attribute_value returns a textual representation of the value of an attribute as reported to a start_element handler or undef if no value is available.

split_message($message, $filename, $open_ent, $error_num, $open_elem)

split_message splits an OpenSP error message into its components, the error or warning message, an optional auxiliary message that provides additional information about the error, like the first occurence of an ID in case of duplicate IDs in a document, each accompanied by line and column numbers relevant to the message, and depending on the parser configuration the open entities for the message, the error number of the message and a list of the current open elements.

It returns a hash reference like

  # this is always present
  primary_message =>
    Number       => 141,       # only if $p->show_error_numbers(1)
    Module       => 554521624, # only if $p->show_error_numbers(1)
    ColumnNumber => 9,
    LineNumber   => 12,
    Severity     => 'E',
    Text         => 'ID "a" already defined'

  # only some messages have an aux_message 
  aux_message =>
    ColumnNumber => 9,
    LineNumber   => 11,
    Text         => 'ID "a" first defined here'

  # iff $p->show_open_elements(1) and there are open elements
  open_elements => 'html body[1] (p[1])',

  # iff $p->show_open_entities(1) and there are open entities
  # other than the document, but the document will be reported
  # if the error is in some other entity
  open_entities => [
    ColumnNumber => 55,
    FileName     => 'example.xhtml',
    EntityName   => 'html',
    LineNumber   => 2
  }, ... ],

This would typically be used like

  sub error
    my $self = shift;
    my $erro = shift;
    my $mess = $erro->{Message};

    # parser is the SGML::Parser::OpenSP
    # object stored in the handler object
    my $loca = $self->{parser}->get_location;
    my $name = $loca->{FileName};

    my $splt = split_message($mess, $name,

    # ...

A more convenient way to access this function is provided by the SGML::Parser::OpenSP module which you can use like

  sub error
    my $self = shift;
    my $erro = shift;

    my $mess = $self->{parser}->split_message($erro);

    # relevant data is now $mess and $erro->{Severity}
    # of which the latter provides more detailed information
    # than $mess->{primary_message}->{Severity}, see the
    # SGML::Parser::OpenSP documentation for details

split_pi splits the data of a processing instructions at the first white space character into two components where white space character is defined in the $WHITESPACE package variable, qr/[\x20\x09\x0d\x0a]/ by default. It returns undef if there is no data to split.

  sub pi
    my $self = shift;
    my $proc = shift;

    my ($target, $data) = split_pi($proc->{Data});

    # ...


  Copyright (c) 2006-2008 Bjoern Hoehrmann <>.
  This module is licensed under the same terms as Perl itself.