|
|
XML::Checker::Parser - an XML::Parser that validates at parse time
use XML::Checker::Parser;
my %expat_options = (KeepCDATA => 1,
Handlers => [ Unparsed => \&my_Unparsed_handler ]);
my $parser = new XML::Checker::Parser (%expat_options);
eval {
local $XML::Checker::FAIL = \&my_fail;
$parser->parsefile ("fail.xml");
};
if ($@) {
# Either XML::Parser (expat) threw an exception or my_fail() died.
... your error handling code here ...
}
# Throws an exception (with die) when an error is encountered, this
# will stop the parsing process.
# Don't die if a warning or info message is encountered, just print a message.
sub my_fail {
my $code = shift;
die XML::Checker::error_string ($code, @_) if $code < 200;
XML::Checker::print_error ($code, @_);
}
XML::Checker::Parser extends the XML::Parser manpage
I hope the example in the SYNOPSIS says it all, just use the XML::Checker::Parser manpage as if it were an XML::Parser. See the XML::Parser manpage for the supported (expat) options.
You can also derive your parser from XML::Checker::Parser instead of from XML::Parser. All you should have to do is replace:
package MyParser; @ISA = qw( XML::Parser );
with:
package MyParser; @ISA = qw( XML::Checker::Parser );
$parser = new XML::Checker::Parser (SkipExternalDTD => 1, SkipInsignifWS => 1);
The constructor takes the same parameters as the XML::Parser manpage with the following additions:
By default, it will try to load external DTDs using LWP. You can disable this by setting SkipExternalDTD to 1. See External DTDs for details.
By default, it will treat insignificant whitespace as regular Char data. By setting SkipInsignifWS to 1, the user Char handler will not be called if insignificant whitespace is encountered. See INSIGNIFICANT_WHITESPACE in the XML::Checker manpage for details.
When calling parsefile() with a URL (instead of a filename) or when loading
external DTDs, we use LWP to download the
remote file. By default it will use a the LWP::UserAgent manpage that is created as follows:
use LWP::UserAgent; $LWP_USER_AGENT = LWP::UserAgent->new; $LWP_USER_AGENT->env_proxy;
Note that env_proxy reads proxy settings from your environment variables, which is what I need to do to get thru our firewall. If you want to use a different LWP::UserAgent, you can either set it globally with:
XML::Checker::Parser::set_LWP_UserAgent ($my_agent);
or, you can specify it for a specific XML::Checker::Parser by passing it to the constructor:
my $parser = new XML::Checker::Parser (LWP_UserAgent => $my_agent);
Currently, LWP is used when the filename (passed to parsefile) starts with one of the following URL schemes: http, https, ftp, wais, gopher, or file (followed by a colon.) If I missed one, please let me know.
The LWP modules are part of libwww-perl which is available at CPAN.
XML::Checker::Parser will try to load and parse external DTDs that are referenced in DOCTYPE definitions unless you set the SkipExternalDTD option to 1 (the default setting is 0.) See CAVEATS for details on what is not supported by XML::Checker::Parser.
the XML::Parser manpage (version 2.27 and up) does a much better job at reading external DTDs, because recently external DTD parsing was added to expat. Make sure you set the the XML::Parser manpage option ParseParamEnt to 1 and the XML::Checker::Parser option SkipExternalDTD to 1. (They can both be set in the XML::Checker::Parser constructor.)
When external DTDs are parsed by XML::Checker::Parser, they are located in the following order:
With the %URI_MAP, which can be set using map_uri. This hash maps external resource ids (like system ID's and public ID's) to full path URI's. It was meant to aid in resolving PUBLIC IDs found in DOCTYPE declarations after the PUBLIC keyword, e.g.
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0//EN">
However, you can also use this to force the XML::Checker manpage to read DTDs from a different URL than was specified (e.g. from the local file system for performance reasons.)
on the Internet, if their system identifier starts with a protocol (like http://...)
on the local disk, if their system identifier starts with a slash (absolute path)
in the SGML_SEARCH_PATH, if their system identifier is a relative file name. It will use @SGML_SEARCH_PATH if it was set with set_sgml_search_path(), or the colon-separated $ENV{SGML_SEARCH_PATH}, or (if that isn't set) the list (``.'', ``$ENV{'HOME'}/.sgml'', ``/usr/lib/sgml'', ``/usr/share/sgml''), which includes the current directory, so it should do the right thing in most cases.
External DTDs with relative file paths are looked up using the @SGML_SEARCH_PATH, which can be set with this method. If @SGML_SEARCH_PATH is never set, it will use the colon-separated $ENV{SGML_SEARCH_PATH} instead. If neither are set it uses the list: ``.'', ``$ENV{'HOME'}/.sgml'', ``/usr/lib/sgml'', ``/usr/share/sgml''.
set_sgml_search_path is a static method.
To define the location of PUBLIC ids, as found in DOCTYPE declarations after the PUBLIC keyword, e.g.
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0//EN">
call this method, e.g.
XML::Checker::Parser::map_uri (
"-//W3C//DTD HTML 4.0//EN" => "file:/user/html.dtd";);
See External DTDs for more info.
XML::Checker::Parser::map_uri is a static method.
You should be able to use setHandlers() just as in the XML::Parser manpage.
(Using setHandlers has not been tested yet.)
XML::Checker::Parser routes the fail handler through XML::Checker::Parser::fail_add_context() before calling your fail handler (i.e. the global fail handler: $XML::Checker::FAIL. See ERROR_HANDLING in the XML::Checker manpage.) It adds the (line, column, byte) information from the XML::Parser manpage to the error context (unless it was the end of the XML document.)
Only the following the XML::Parser manpage handlers are currently routed through the XML::Checker manpage: Init, Final, Char, Start, End, Element, Attlist, Doctype, Unparsed, Notation.
When using XML::Checker::Parser to parse external DTDs (i.e. with SkipExternalDTD => 0), expect trouble when your external DTD contains parameter entities inside declarations or conditional sections. The external DTD should probably have the same encoding as the orignal XML document.
Send bug reports, hints, tips, suggestions to Enno Derksen at <enno@att.com>.
the XML::Checker manpage (SEE_ALSO in the XML::Checker manpage), the XML::Parser manpage