=encoding utf8

=head1 NAME

XML::Compile::Cache - Cache compiled XML translators


   is a XML::Compile::Schema
   is a XML::Compile


 my $cache = XML::Compile::Cache->new(...);

 $cache->declare('READER',  $type,  @options);
 $cache->declare(RW     => \@types, @options);
 $cache->declare(WRITER =>  $type, \@options);


 # get the cached code ref for the reader
 my $reader = $cache->reader($type, @opts);
 use Data::Dumper;
 print Dumper $reader->($xml);

 # get the cached code ref for the writer, and use it
 my $doc = XML::LibXML::Document->new('1.0', 'UTF-8');
 my $xml = $cache->writer($type)->($doc, $perl);
 print $xml->toString(1);

 # use the base-class uncached, the XML::Compile::Schema
 my $do = $cache->compile(READER => $type, @opts);


Extends L<"DESCRIPTION" in XML::Compile::Schema|XML::Compile::Schema/"DESCRIPTION">.
=head1 METHODS

Extends L<"METHODS" in XML::Compile::Schema|XML::Compile::Schema/"METHODS">.
=head2 Constructors

Extends L<"Constructors" in XML::Compile::Schema|XML::Compile::Schema/"Constructors">.
=over 4

=item XML::Compile::Cache-E<gt>B<new>( [$xml], %options )

 -Option            --Defined in          --Default
  allow_undeclared                          <false>
  any_element                               'ATTEMPT'
  block_namespace     XML::Compile::Schema  []
  hook                XML::Compile::Schema  undef
  hooks               XML::Compile::Schema  []
  ignore_unused_tags  XML::Compile::Schema  <false>
  key_rewrite         XML::Compile::Schema  []
  opts_readers                              []
  opts_rw                                   []
  opts_writers                              []
  parser_options      XML::Compile          <many>
  prefixes                                  <smart>
  schema_dirs         XML::Compile          undef
  typemap                                   {}
  xsi_type                                  {}

=over 2

=item allow_undeclared => BOOLEAN

When true, you may call the reader or writer with types which were
not registered with L<declare()|XML::Compile::Cache/"Administration">.  In that case, the reader or
writer may also get options passed for the compiler, as long as
they are consistent over each use of the type.

=item any_element => CODE|'TAKE_ALL'|'SKIP_ALL'|'ATTEMPT'|'SLOPPY'

See L<anyElement()|XML::Compile::Cache/"Accessors">.

[1.02] the default is to ATTEMPT compiling any handlers automatically.
Before version 1.02, the default was to SKIP_ALL elements which would
match the occurs and namespace restrictions of the any specification.
However, that fails for reperative blocks (for instance, it fails for
an choice which may occur unbounded times)

=item block_namespace => NAMESPACE|TYPE|HASH|CODE|ARRAY

=item hook => $hook|ARRAY

=item hooks => ARRAY

=item ignore_unused_tags => BOOLEAN|REGEXP

=item key_rewrite => HASH|CODE|ARRAY

=item opts_readers => HASH|ARRAY-of-PAIRS

=item opts_rw => HASH|ARRAY-of-PAIRS

Options added to both READERs and WRITERS.  Options which are passed
with L<declare()|XML::Compile::Cache/"Administration"> and C<opts_readers> or C<opts_writers> will overrule
these.  See L<addCompileOptions()|XML::Compile::Cache/"Compilers">.

=item opts_writers => HASH|ARRAY-of-PAIRS

=item parser_options => HASH|ARRAY

=item prefixes => HASH|ARRAY-of-PAIRS

Define prefix name to name-space mappings.  Passed to L<compile(prefixes)|XML::Compile::Schema/"Compilers">
for each reader and writer, but also used to permit L<findName()|XML::Compile::Cache/"Administration"> to
accept types which use a prefix.

Specify an ARRAY of (prefix, name-space) pairs, or a HASH which maps
name-spaces to prefixes (HASH order is reversed from ARRAY order!)  When
you wish to collect the results, like usage counts, of the translation
processing, you will need to specify a HASH.

 prefixes => [ mine => $myns, your => $yourns ]
 prefixes => { $myns => 'mine', $yourns => 'your' }

 # the previous is short for:
 prefixes => { $myns => [ uri => $myns, prefix => 'mine', used => 0 ]
             , $yourns => [ uri => $yourns, prefix => 'your', ...] }

=item schema_dirs => $directory|ARRAY-OF-directories

=item typemap => HASH|ARRAY

=item xsi_type => HASH|ARRAY



=head2 Accessors

Extends L<"Accessors" in XML::Compile::Schema|XML::Compile::Schema/"Accessors">.
=over 4

=item $obj-E<gt>B<addHook>($hook|LIST|undef)

Inherited, see L<XML::Compile::Schema/"Accessors">

=item $obj-E<gt>B<addHooks>( $hook, [$hook, ...] )

Inherited, see L<XML::Compile::Schema/"Accessors">

=item $obj-E<gt>B<addKeyRewrite>($predef|CODE|HASH, ...)

Inherited, see L<XML::Compile::Schema/"Accessors">

=item $obj-E<gt>B<addSchemaDirs>(@directories|$filename)

=item XML::Compile::Cache-E<gt>B<addSchemaDirs>(@directories|$filename)

Inherited, see L<XML::Compile/"Accessors">

=item $obj-E<gt>B<addSchemas>($xml, %options)

Inherited, see L<XML::Compile::Schema/"Accessors">

=item $obj-E<gt>B<addTypemap>(PAIR)

Inherited, see L<XML::Compile::Schema/"Accessors">

=item $obj-E<gt>B<addTypemaps>(PAIRS)

Inherited, see L<XML::Compile::Schema/"Accessors">

=item $obj-E<gt>B<addXsiType>( [HASH|ARRAY|LIST] )

[1.01] add global xsi_type declarations.  Returns the xsiType set.
The ARRAY or LIST contains pairs, just like the HASH.

The value component can be 'AUTO' to automatically detect the C<xsi:type>
extensions.  This does only work for complex types.

=item $obj-E<gt>B<allowUndeclared>( [BOOLEAN] )

Whether it is permitted to create readers and writers which are not
declared cleanly.

=item $obj-E<gt>B<anyElement>('ATTEMPT'|'SLOPPY'|'SKIP_ALL'|'TAKE_ALL'|CODE)

[as method since 0.99] How to process ANY elements, see also

Reader: C<ATTEMPT> will convert all any elements, applying the reader for
each element found. When an element is not found in a schema, it will
be included as XML::LibXML::Element node.

[0.93] Reader: With C<SLOPPY>, first automatic typed conversion is
attempted. But is the type is not known, L<XML::LibXML::Simple::XMLin()|XML::LibXML::Simple/"Translators">
is called to the resque.

=item $obj-E<gt>B<blockNamespace>($ns|$type|HASH|CODE|ARRAY)

Inherited, see L<XML::Compile::Schema/"Accessors">

=item $obj-E<gt>B<hooks>( [<'READER'|'WRITER'>] )

Inherited, see L<XML::Compile::Schema/"Accessors">

=item $obj-E<gt>B<typemap>( [HASH|ARRAY|PAIRS] )

[0.98] Add global knowledge on typemaps.  Returns the typemap.

=item $obj-E<gt>B<useSchema>( $schema, [$schema, ...] )

Inherited, see L<XML::Compile::Schema/"Accessors">


=head2 Prefix management

The cache layer on top of L<XML::Compile::Schema|XML::Compile::Schema> adds smart use of
prefixes.  Of course, smartness comes with a small performance cost,
but the code gets much cleaner.

=over 4

=item $obj-E<gt>B<addNicePrefix>(BASE, NAMESPACE)

[1.03] Register NAMESPACE -if not yet defined- with prefix name BASE.  When
that prefix name is already in use for some other namespace, BASE followed
by a number are attempted (starting with 01).  The prefix is returned.

When the BASE already ends on a number, that number will get counted.


  my $prefix = $schema->addNicePrefix('call', $myns);
  # $prefix now can be call, call01, call02 etc

=item $obj-E<gt>B<addPrefixes>( [PAIRS|ARRAY|HASH] )

The X::C logic does auto-detect prefix/namespaces combinations from
the XML, but does not search extensively for namespace declarations.
Also, sometimes the same namespace is used with different prefixes.
Sometimes, the same prefix is used for different namesapces.  To complete
the list, or control the actual prefix being used, you explicitly declare

The B<best way> to add prefixes is via L<new(prefixes)|XML::Compile::Cache/"Constructors">, which will give
your names preference over the names found in the schema's which get loaded.
For instance, use C<< ::WSDL->new(prefixes => [ $prefix => $ns ] >>

[0.995] Returns the HASH with prefix to name-space translations.  You should
not modify the returned HASH: new PAIRS of prefix to namespace relations
can be passed as arguments.

[0.14] If a name-space appears for the second time, then the new
prefix will be recognized by L<findName()|XML::Compile::Cache/"Administration">, but not used in the output.
When the prefix already exists for a different namespace, then an error
will be casted.

[0.90] You may also provide an ARRAY of pairs or a HASH.

=item $obj-E<gt>B<learnPrefixes>($node)

[0.993] Take all the prefixes defined in the $node, and XML::LibXML::Element.
This is not recursive: only on those defined at the top $node.

=item $obj-E<gt>B<prefix>($prefix)

Lookup a prefix definition.  This returns a HASH with namespace info.

=item $obj-E<gt>B<prefixFor>($uri)

Lookup the preferred prefix for the $uri.

=item $obj-E<gt>B<prefixed>( $type|<$ns,$local> )

Translate the fully qualified $type into a prefixed version.  Will produce
undef if the namespace is unknown.

[0.993] When your $type is not in packed form, you can specify a namespace
and $local type name as separate arguments.


   print $schema->prefixed($type) || $type;
   print $schema->prefixed($ns, $local);

=item $obj-E<gt>B<prefixes>( [$params] )

Return prefixes table.  The $params are deprecated since [0.995], see
L<addPrefixes()|XML::Compile::Cache/"Prefix management">.


=head2 Compilers

The name of this module refers to its power to administer compiled
XML encoders (writers) and decoders (readers).  This means that
your program only need to pass on a ::Cache object (for instance
a XML::Compile::WSDL11, not a CODE reference for each compiled

Extends L<"Compilers" in XML::Compile::Schema|XML::Compile::Schema/"Compilers">.
=over 4

=item $obj-E<gt>B<addCompileOptions>( ['READERS'|'WRITERS'|'RW'], %options )

[0.99] You may provide global compile options with L<new(opts_rw)|XML::Compile::Cache/"Constructors">,
C<opts_readers> and C<opts_writers>, but also later using this method.

=item $obj-E<gt>B<compile>( <'READER'|'WRITER'>, $type, %options )

Inherited, see L<XML::Compile::Schema/"Compilers">

=item $obj-E<gt>B<compileAll>( ['READERS'|'WRITERS'|'RW', [$ns]] )

Compile all the declared readers and writers with the default 'RW').  You may
also select to pre-compile only the READERS or only the WRITERS.  The
selection can be limited further by specifying a $ns.

By default, the processors are only compiled when used.  This method is
especially useful in a B<daemon process>, where preparations can take as
much time as they want to... and running should be as fast as possible.

=item $obj-E<gt>B<compileType>( <'READER'|'WRITER'>, $type, %options )

Inherited, see L<XML::Compile::Schema/"Compilers">

=item $obj-E<gt>B<dataToXML>($node|REF-XML|XML-STRING|$filename|$fh|$known)

=item XML::Compile::Cache-E<gt>B<dataToXML>($node|REF-XML|XML-STRING|$filename|$fh|$known)

Inherited, see L<XML::Compile/"Compilers">

=item $obj-E<gt>B<initParser>(%options)

=item XML::Compile::Cache-E<gt>B<initParser>(%options)

Inherited, see L<XML::Compile/"Compilers">

=item $obj-E<gt>B<reader>($type|$name, %options)

Returns the reader CODE for the $type or $name (see L<findName()|XML::Compile::Cache/"Administration">).
%options are only permitted if L<new(allow_undeclared)|XML::Compile::Cache/"Constructors"> is true, and the
same as the previous call to this method.

The reader will be compiled the first time that it is used, and that
same CODE reference will be returned each next request for the same
$type.  Call L<compileAll()|XML::Compile::Cache/"Compilers"> to have all readers compiled by force.

 -Option --Default
  is_type  <false>

=over 2

=item is_type => BOOLEAN

[1.03] use L<compileType()|XML::Compile::Schema/"Compilers"> with the given element, to replace L<compile()|XML::Compile::Schema/"Compilers">
You probably want an additional C<element> parameter.



  my $schema = XML::Compile::Cache->new(\@xsd,
     prefixes => [ gml => $GML_NAMESPACE ] );
  my $data   = $schema->reader('gml:members')->($xml);

  my $getmem = $schema->reader('gml:members');
  my $data   = $getmem->($xml);

=item $obj-E<gt>B<template>( <'XML'|'PERL'|'TREE'>, $element, %options )

Inherited, see L<XML::Compile::Schema/"Compilers">

=item $obj-E<gt>B<writer>($type|$name)

Returns the writer CODE for the $type or $name (see L<findName()|XML::Compile::Cache/"Administration">).
OPTIONS are only permitted if L<new(allow_undeclared)|XML::Compile::Cache/"Constructors"> is true, and the
same as the previous call to this method.

The writer will be compiled the first time that it is used, and that
same CODE reference will be returned each next request for the same

 -Option --Default
  is_type  <false>

=over 2

=item is_type => BOOLEAN

[1.03] use L<compileType()|XML::Compile::Schema/"Compilers"> with the given element, to replace L<compile()|XML::Compile::Schema/"Compilers">
You probably want an additional C<element> parameter.



  my $xml = $cache->writer('gml:members')->($doc, $data);

  my $doc = XML::LibXML::Document->new('1.0', 'UTF-8');
  my $wr  = $cache->writer('gml:members');
  my $xml = $wr->($doc, $data);
  print $doc->toString(1);


=head2 Administration

Extends L<"Administration" in XML::Compile::Schema|XML::Compile::Schema/"Administration">.
=over 4

=item $obj-E<gt>B<declare>( <'READER'|'WRITER'|'RW'>, <$type|ARRAY>, %options )

Register that the indicated $type (or ARRAY of them) may be used, and needs to
be translated with the %options (either specified as ARRAY or PAIRS).
Specify whether it may get used as READER, WRITER, or both (RW).  If the
READER and WRITER need different options, then you need to declare them
separately; in that case you cannot use RW.

The $type should be understood by L<findName()|XML::Compile::Cache/"Administration">, so may be prefixed.


  $cache->declare(READER => 'pref:count', sloppy_integers => 1)
        ->declare(RW     => '{myns}mylocal');

  $cache->declare(WRITER => [ 'xsd:int', '{http://}aap' ]);

=item $obj-E<gt>B<doesExtend>($exttype, $basetype)

Inherited, see L<XML::Compile::Schema/"Administration">

=item $obj-E<gt>B<elements>()

Inherited, see L<XML::Compile::Schema/"Administration">

=item $obj-E<gt>B<findName>($name)

Translate the $name specification into a schema defined full type.
The $name can be a full type (like '{namespace}localname', usually
created with L<XML::Compile::Util::pack_type()|XML::Compile::Util/"Packing">) or a prefixed type
(like 'myns:localname', where C<myns> is defined via L<new(prefixes)|XML::Compile::Cache/"Constructors">
or L<prefixes()|XML::Compile::Cache/"Prefix management">).

When the form is 'myns:' (so without local name), the namespace uri is

example: of findName()

  $schema->addPrefixes(pre => 'http://namespace');

  my $type = $schema->findName('pre:name');
  print $type;   # {http://namespace}name

  my $ns   = $schema->findName('pre:');
  print $ns;     # http://namespace

  my $type = $schema->findName('{somens}name');
  print $type;   # {somens}name    [a no-op]

=item $obj-E<gt>B<findSchemaFile>($filename)

=item XML::Compile::Cache-E<gt>B<findSchemaFile>($filename)

Inherited, see L<XML::Compile/"Administration">

=item $obj-E<gt>B<importDefinitions>($xmldata, %options)

Inherited, see L<XML::Compile::Schema/"Administration">

=item $obj-E<gt>B<knownNamespace>($ns|PAIRS)

=item XML::Compile::Cache-E<gt>B<knownNamespace>($ns|PAIRS)

Inherited, see L<XML::Compile/"Administration">

=item $obj-E<gt>B<namespaces>()

Inherited, see L<XML::Compile::Schema/"Administration">

=item $obj-E<gt>B<printIndex>( [$fh], %options )

 -Option       --Default
  show_declared  <true>

=over 2

=item show_declared => BOOLEAN

Add an indicator to each line, about whether readers and writers are
declare for the type.  Declared readers and writers will show flags
C<r> and C<w> respectively.  Compiled readers and writers carry a
C<R> and/or C<W>.


=item $obj-E<gt>B<types>()

Inherited, see L<XML::Compile::Schema/"Administration">

=item $obj-E<gt>B<walkTree>($node, CODE)

Inherited, see L<XML::Compile/"Administration">


=head1 DETAILS

Extends L<"DETAILS" in XML::Compile::Schema|XML::Compile::Schema/"DETAILS">.

C<XML::Compile::Cache> is the smart brother of L<XML::Compile::Schema|XML::Compile::Schema>;
it keeps track of your compiled readers and writers, and also helps
you administer the parameters to handle compilation.  Besides, it
lat you use easy prefixes instead of full namespaces.

With L<XML::Compile::Schema::compile()|XML::Compile::Schema/"Compilers"> (defined in the SUPER class of
this module) you can construct translators from XML to Perl and back.
These translators are code references, which are "expensive" to create,
but "cheap" in use; call them as often as you want.  This module helps
you administer them.

When the schemas grow larger, it gets harder to see which code reference
have already be created and which not. And, these code references need
compile options which you do not want to distribute over your whole
program.  Finally, in a daemon application, you do not want to create
the translators when used (which can be in every client again), but once
during the initiation of the daemon.

One of the most important contributions to the compile management, is
the addition of smart prefix handling. This means that you can use
prefixed names in stead of full types, often created with

=head1 SEE ALSO

This module is part of XML-Compile-Cache distribution version 1.06,
built on March 04, 2018. Website: F<http://perl.overmeer.net/xml-compile/>

=head1 LICENSE

Copyrights 2008-2018 by [Mark Overmeer]. For other contributors see ChangeLog.

This program is free software; you can redistribute it and/or modify it
under the same terms as Perl itself.
See F<http://dev.perl.org/licenses/>