XMLin()¶

The XMLin() function takes the dtd and a string as arguments and returns a hash reference containing the data.

XMLinfile()¶

This function works like XMLin(), but takes a filename or a file handle as second argument.

XMLout()¶

XMLout() provides the reverse operation to XMLin(), it takes a dtd and a hash reference as arguments and returns an XML string.

XMLoutfile() This function works like XMLout(), but takes a filename or a file handle as third argument.¶

simple attributes and elements¶

The very simple example for a dtd is:

    $dtd = [ 'user' =>
                 'login',
                 'password',
           ];

This dtd will accept/create XML like:

    <user login="foo" password="bar" />

XMLin doesn't care if "login" or "password" are attributes or elements, so

    <user>
      <login>foo</login>
      <password>bar</password>
    </user>

is also valid input (but doesn't get re-created by XMLout()).

multiple elements of the same name¶

If an element may appear multiple times, it must be declared as an array in the dtd:

    $dtd = [ 'user' =>
                 'login',
                 [ 'favorite_fruits' ],
           ];

XMLin will create an array reference as value in this case, even if the xml data contains only one element. Valid XML looks like:

    <user login="foo">
      <favorite_fruits>apple</favorite_fruits>
      <favorite_fruits>peach</favorite_fruits>
    </user>

As attributes may not appear multiple times, XMLout will create elements for this case. Note also that all attributes must come before the first element, thus the first array in the dtd ends the attribute list. As an example, the following dtd

    $dtd = [ 'user' =>
                 'login',
                 [ 'favorite_fruits' ],
                 'password',
           ];

will create xml like:

    <user login="foo">
      <favorite_fruits>apple</favorite_fruits>
      <favorite_fruits>peach</favorite_fruits>
      <password>bar</password>
    </user>

"login" is translated to an attribute and "password" to an element.

You can use an empty array reference to force the end of the attribute list, e.g.:

    $dtd = [ 'user' =>
                 [],
                 'login',
                 'password',
           ];

will translate to

    <user>
      <login>foo</login>
      <password>bar</password>
    </user>

instead of

    <user login="foo" password="bar" />

sub-elements¶

sub-elements are elements that also contain attributes or other elements. They are specified in the dtd as arrays with more than one element. Here is an example:

    $dtd = [ 'user' =>
                 'login',
                 [ 'address' =>
                     'street',
                     'city',
                 ],
           ];

Valid xml for this dtd looks like:

    <user login="foo">
      <address street="broadway 7" city="new york" />
    </user>

It is sometimes useful to specify such dtds in multiple steps:

    $addressdtd = [ 'address' =>
                         'street',
                         'city',
                  ];
    $dtd = [ 'user' =>
                 'login',
                 $addressdtd,
           ];

multiple sub-elements with the same name¶

As with simple elements, one can allow sub-elements to occur multiple times. XMLin() creates an array of hash references in this case. The dtd specification uses an array reference to an array for this case, for example:

    $dtd = [ 'user' =>
                 'login',
                 [[ 'address' =>
                     'street',
                     'city',
                 ]],
           ];
Or, with the $addressdtd definition used in the previous example:
    $dtd = [ 'user' =>
                 'login',
                 [ $addressdtd ],
           ];

Accepted XML is:

    <user login="foo">
      <address street="broadway 7" city="new york" />
      <address street="rural road 12" city="tempe" />
    </user>

the _content pseudo-element¶

All of the non-whitespace parts between elements get collected into a single "_content" element. As example,

    <user login="foo">
      <address street="broadway 7" city="new york"/>hello
      <address street="rural road 12" city="tempe"/>world
    </user>

would set the _content element to "hello\n world" (the dtd must allow a _content element, of course). If the dtd is

    $dtd = [ 'user' =>
                 'login',
                 [ $addressdtd ],
                 '_content',
           ];

the xml string created by XMLout() will be:

    <user login="foo">
      <address street="broadway 7" city="new york" />
      <address street="rural road 12" city="tempe" />
      hello
      world    
    </user>

The exact input cannot be re-created, as the positions and the fragmentation of the content data is lost.